H2 Console: command line properties are now longer stored in the properties file. They are now only used for the current process, except if they are explicitly saved.
Identifiers with a digit and then a dollar sign now work. Example: A1$B.
@changelog_1003_li
MS SQL Server compatibility: support for linked tables with NVARCHAR, NCHAR, NCLOB, and LONGNVARCHAR.
@changelog_1004_li
Android: Workaround for a problem when using read-only databases in zip files (skip seems to be implemented incorrectly on the Android system).
@changelog_1005_li
Calling execute() or prepareStatement() with null as the SQL statement now throws an exception.
@changelog_1006_li
Benchmark: the number of executed statements was incorrect. The H2 database was loaded at the beginning of the test to collect results, now it is loaded at the very end. Thanks to Fred Toussi from HSQLDB for reporting those problems. However the changed do not affect the relative performance.
@changelog_1007_li
H2 Console: command line settings are now longer stored in the properties file. They are now only used for the current process, except if they are explicitly saved.
@changelog_1008_li
Cache: support for a second level soft-references cache. To enable it, append ;CACHE_TYPE=SOFT_LRU (or SOFT_TQ) to the database URL, or set the system property h2.cacheTypeDefault to "SOFT_LRU" / "SOFT_TQ". Enabling the second level cache reduces performance for small databases, but speeds up large databases. It makes sense to use it if the available memory size is unknown. Thanks a lot to Jan Kotek!
@changelog_1004_h2
@changelog_1009_h2
Version 1.1.112 (2009-05-01)
@changelog_1005_li
@changelog_1010_li
JdbcPreparedStatement.toString() could throw a NullPointerException.
@changelog_1006_li
@changelog_1011_li
EclipseLink: Added H2Platform.supportsIdentity().
@changelog_1007_li
@changelog_1012_li
Connection pool: the default login timeout is now 5 minutes.
@changelog_1008_li
@changelog_1013_li
After truncating tables, opening large databases could become slow because indexes were always re-built unnecessarily when opening.
@changelog_1009_li
@changelog_1014_li
More bugs in the server-less multi-connection mode have been fixed: Sometimes parameters of prepared statements were lost when a reconnecting. Concurrent read operations were slow. To improve performance, executeQuery(..) must be used for queries (execute(..) switches to the write mode, which is slow).
@changelog_1010_li
@changelog_1015_li
GROUP BY queries with a self-join (join to the same table) that were grouped by columns with indexes returned the wrong result in some cases.
@changelog_1011_li
@changelog_1016_li
Improved error message when the .lock.db file modification time is in the future.
@changelog_1012_li
@changelog_1017_li
The MERGE statement now returns 0 as the generated key if the row was updated.
@changelog_1013_li
@changelog_1018_li
Running code coverage is now automated.
@changelog_1014_li
@changelog_1019_li
A file system implementation can now be registered using FileSystem.register.
@changelog_1015_li
@changelog_1020_li
The database file system is no longer included in the jar file, it moved to the test section.
@changelog_1016_h2
@changelog_1021_h2
Version 1.1.111 (2009-04-10)
@changelog_1017_li
@changelog_1022_li
In-memory databases can now run inside the Google App Engine.
@changelog_1018_li
@changelog_1023_li
Queries that are ordered by an indexed column returned no rows in certain cases (if all rows were deleted from the table previously, and there is a low number of rows in the table, and when not using other conditions, and when using the default b tree index).
@changelog_1019_li
@changelog_1024_li
The wrong exception was thrown when using unquoted text for the SQL statements COMMENT, ALTER USER, and SET PASSWORD.
@changelog_1020_li
@changelog_1025_li
The built-in connection pool did not roll back transactions and enable autocommit enabled after closing a connection.
@changelog_1021_li
@changelog_1026_li
Sometimes a StackOverflow occurred when checking for deadlock. See also http://code.google.com/p/h2database/issues/detail?id=61
@changelog_1022_li
@changelog_1027_li
The Shell tool no longer truncates results with only one column, and displays a message if data was truncated.
@changelog_1023_h2
@changelog_1028_h2
Version 1.1.110 (2009-04-03)
@changelog_1024_li
@changelog_1029_li
Support for not persistent in-memory tables in regular (persistent) databases using CREATE MEMORY TABLE(..) NOT PERSISTENT. Thanks a lot to Sergi Vladykin for the patch!
@changelog_1025_li
@changelog_1030_li
The H2 Console trimmed the password (removed leading and trailing spaces). This is no longer the case, to support encrypted H2 database with an empty user password.
@changelog_1026_li
@changelog_1031_li
The data type of a SUBSTRING method was wrong.
@changelog_1027_li
@changelog_1032_li
ResultSet.findColumn and get methods with column label parameters now also check for matching column names (like most databases except MySQL).
@changelog_1028_li
@changelog_1033_li
H2 Console: the browser system property now supports a list of arguments. Example: java -Dh2.browser="open,-a,Safari,%url" ...
@changelog_1029_li
@changelog_1034_li
Improved Javadoc navigation (similar to Scaladoc).
@changelog_1030_li
@changelog_1035_li
H2 Console: auto-complete of identifiers did not work correctly for H2 databases in MySQL mode.
@changelog_1031_li
@changelog_1036_li
DISTINCT and GROUP BY on a CLOB column was broken.
@changelog_1032_li
@changelog_1037_li
The FTP server moved to the tools section and is no longer included in the h2*.jar file.
@changelog_1033_li
@changelog_1038_li
Improved error message for unsupported features: now the message says what exactly is not supported.
@changelog_1034_li
@changelog_1039_li
Improved OSGi support.
@changelog_1035_li
@changelog_1040_li
Some internal caches did not use the LRU mechanism. Fixed (LOB file list, optimizer cost cache, trace system, view indexes, collection keys, compressed in-memory file system).
@changelog_1036_li
@changelog_1041_li
The API of the tools changed a bit (each tool now returns an exit code).
@changelog_1037_li
@changelog_1042_li
Command line help of the tools now match the javadocs. The build converts the javadocs to a resource that is read by the tool at runtime. This should not have an effect on using the database, but it reduces duplicate and out-of-sync documentation.
@changelog_1038_li
@changelog_1043_li
CREATE TABLE: improved compatibility (support for UNIQUE NOT NULL).
@changelog_1039_li
@changelog_1044_li
DatabaseMetaData.getSQLKeywords now returns the correct list.
@changelog_1040_li
@changelog_1045_li
Deterministic user defined functions did not work when the parameter was a column. Fixed.
@changelog_1041_li
@changelog_1046_li
JdbcConnectionPool.setLoginTimeout with 0 now uses the default timeout.
@changelog_1042_li
@changelog_1047_li
Creating a JdbcConnectionPool has been simplified a bit.
@changelog_1043_li
@changelog_1048_li
The built-in connection pool did not re-use connections. Getting a connection using the built-in JdbcConnectionPool is now about 70 times faster than opening connections using DriverManager.getConnection.
@changelog_1044_li
@changelog_1049_li
More bugs in the server-less multi-connection mode have been fixed: If a process terminated while writing, other open connections were blocked. If two processes were writing to the database, sometimes the database was corrupt after closing.
@changelog_1045_li
@changelog_1050_li
Linked tables to SQLite database can now be created.
@changelog_1046_li
@changelog_1051_li
Nested IN(IN(...)) didn't work.
@changelog_1047_li
@changelog_1052_li
NIO storage: the nio: prefix was using memory mapped files instead of FileChannel.
@changelog_1048_h2
@changelog_1053_h2
Version 1.1.109 (2009-03-14)
@changelog_1049_li
@changelog_1054_li
The optimization for IN(...) is now only used if comparing a column with an index.
@changelog_1050_li
@changelog_1055_li
User defined functions can now be deterministic (see CREATE ALIAS documentation).
@changelog_1051_li
@changelog_1056_li
Multiple nested queries in the FROM clause with parameters did not always work.
@changelog_1052_li
@changelog_1057_li
When converting CLOB to BINARY, each character resulted in one byte. Now, the text is parsed as a hex as when converting VARCHAR.
@changelog_1053_li
@changelog_1058_li
New experimental NIO storage mechanism with both FileChannel and memory mapped files. To use it, use the file name prefix nio: or nioMapped: as in jdbc:h2:nio:~/test. So far it looks like NIO storage is faster on Mac OS but slower on some Windows systems. Thanks a lot to Jan Kotek for the patch!
@changelog_1054_li
@changelog_1059_li
The functions BITOR, BITAND, BITXOR, and MOD now accept and return BIGINT instead of INT.
@changelog_1055_li
@changelog_1060_li
Could not use the same linked table multiple times in the same query.
@changelog_1056_li
@changelog_1061_li
Bugs in the server-less multi-connection mode have been fixed.
@changelog_1057_li
@changelog_1062_li
Column names could not be named "UNIQUE" (with the quotes).
@changelog_1058_li
@changelog_1063_li
New system function TRANSACTION_ID() to get the current transaction identifier for a session.
@changelog_1059_h2
@changelog_1064_h2
Version 1.1.108 (2009-02-28)
@changelog_1060_li
@changelog_1065_li
When the shutdown hook closed the database, the last log file was deleted too early. This could cause uncommitted changes to be persisted. In some cases, this could cause data corruption.
@changelog_1061_li
@changelog_1066_li
JdbcConnectionPool: it was possible to set a negative connection pool size.
@changelog_1062_li
@changelog_1067_li
Fulltext search did not support table names with a backslash.
@changelog_1063_li
@changelog_1068_li
The internal IntArray class did not work correctly when initialized with a zero length array.
@changelog_1064_li
@changelog_1069_li
The H2 Console web application (war file) did only support ASCII characters. Now UTF-8 is supported.
@changelog_1065_li
@changelog_1070_li
DATEADD does no longer require that the argument is a timestamp.
@changelog_1066_li
@changelog_1071_li
The database file locking mechanism didn't work correctly on Mac OS.
@changelog_1067_li
@changelog_1072_li
Some built-in functions reported the wrong precision, scale, and display size.
@changelog_1068_li
@changelog_1073_li
MySQL compatibility for CREATE TABLE is improved (UNSIGNED, KEY).
@changelog_1069_li
@changelog_1074_li
Recovery did not work if there were more than 255 lobs stored as files.
@changelog_1070_li
@changelog_1075_li
New experimental mode to support multiple read-write connections without starting a server. To enable this mode, append ;FILE_LOCK=SERIALIZED;OPEN_NEW=TRUE to the database URL. Don't expect high performance when multiple concurrent writers.
@changelog_1071_li
@changelog_1076_li
In a web application, the database classes are not unloaded if a connection is open. This may cause out of memory when re-deploying a web application. The DbStarter is changed to close all connections to the configured database (by executing SHUTDOWN).
@changelog_1072_li
@changelog_1077_li
The WebServlet did not close the database when un-deploying the web application.
@changelog_1073_li
@changelog_1078_li
The exception message of failed INSERT or MERGE statements now includes all values and the row number.
@changelog_1074_li
@changelog_1079_li
If opening a database failed with an out of memory exception, some files were not closed.
@changelog_1075_li
@changelog_1080_li
Optimizer: the expected runtime calculation was incorrect. The fixed calculation should give slightly better query plans when using many joins.
@changelog_1076_li
@changelog_1081_li
Improved exception message when connecting to a just started server fails.
@changelog_1077_li
@changelog_1082_li
Connection.isValid is a bit faster.
@changelog_1078_li
@changelog_1083_li
H2 Console: The autocomplete feature has been improved a bit. It can now better parse conditions.
@changelog_1079_li
@changelog_1084_li
When restarting a web application in Tomcat, an exception was thrown sometimes. In most cases this was a NullPointerException. A workaround in H2 has been implemented. The root cause of the problem is now documented in the FAQ: Tomcat sets all static fields (final or non-final) to null when unloading a web application. A workaround is to put the h2.jar in the lib directory, or set the system property org.apache.catalina.loader.WebappClassLoader.ENABLE_CLEAR_REFERENCES to false.
@changelog_1080_h2
@changelog_1085_h2
Version 1.1.107 (2009-01-24)
@changelog_1081_li
@changelog_1086_li
Some DatabaseMetaData operations did not work for non-admin users for versions 1.1.x.
@changelog_1082_li
@changelog_1087_li
The MySQL compatibility extension fromUnixTime now used the English locale.
@changelog_1083_li
@changelog_1088_li
When using LOG=2 and repeatedly updating the last row rows of a table, the index file grew quickly.
@changelog_1084_li
@changelog_1089_li
In versions 1.1.105 and 1.1.106, encrypted script files of earlier versions could not be processed. This is now again possible. The problem was that such script files were stored in a special format (STORAGE=TEXT) but support for this format was removed in version 1.1.105.
@changelog_1085_li
@changelog_1090_li
Enabling the trace mechanism by creating a specially named file is no longer supported.
@changelog_1086_h2
@changelog_1091_h2
Version 1.1.106 (2009-01-04)
@changelog_1087_li
@changelog_1092_li
Statement.setQueryTimeout did not work correctly for some statements.
@changelog_1088_li
@changelog_1093_li
CREATE DOMAIN: Built-in data types can now only be changed if no tables exist.
@changelog_1089_li
@changelog_1094_li
Linked tables: a workaround for Oracle DATE columns has been implemented.
@changelog_1090_li
@changelog_1095_li
DatabaseMetaData.getPrimaryKeys: The column PK_NAME now contains the constraint name instead of the index name (compatibility for PostgreSQL and Derby).
@changelog_1091_li
@changelog_1096_li
Using IN(..) inside a IN(SELECT..) did not always work.
@changelog_1092_li
@changelog_1097_li
Views with IN(..) that used a view itself did not work.
@changelog_1093_li
@changelog_1098_li
Union queries with LIMIT or ORDER BY that are used in a view or subquery did not work.
@changelog_1094_li
@changelog_1099_li
The license change a bit: so far the license was modified to say 'Swiss law'. This is now changed back to the original 'US law'. This was requested by a user, and I don't see a problem.
@changelog_1095_li
@changelog_1100_li
Constraints for local temporary tables now session scoped. So far they were global. Thanks a lot to Eric Faulhaber for finding and fixing this problem!
@changelog_1096_li
@changelog_1101_li
When using the auto-server mode, and if the lock file was modified in the future, the wrong exception was thrown ('Connection is broken' instead of 'Error opening database: Lock file modified in the future').
@changelog_1097_h2
@changelog_1102_h2
Version 1.1.105 (2008-12-19)
@changelog_1098_li
@changelog_1103_li
The setting STORAGE=TEXT is no longer supported.
@changelog_1099_li
@changelog_1104_li
Deleting a database using the tool DeleteDbFiles deleted LOB files of other databases in the same directory.
@changelog_1100_li
@changelog_1105_li
When used in a subquery, LIKE and IN(..) did not work correctly sometimes.
@changelog_1101_li
@changelog_1106_li
The fulltext search documentation has been improved.
@changelog_1102_li
@changelog_1107_li
ARRAY_GET returned the wrong data type (ARRAY). Now it returns VARCHAR.
@changelog_1103_li
@changelog_1108_li
Natural join: the joined columns are not repeated any more when using SELECT *.
@changelog_1104_li
@changelog_1109_li
User defined aggregate functions: the method getType expected internal data types instead of SQL types.
@changelog_1105_li
@changelog_1110_li
User defined aggregate functions did not work if there was no group by expression.
@changelog_1106_li
@changelog_1111_li
MySQL compatibility: support for := assignment as in @sum:=@sum+x
@changelog_1107_li
@changelog_1112_li
INSERT INTO TEST(SELECT * FROM TEST) is now supported.
@changelog_1108_li
@changelog_1113_li
Each session threw an invisible exception when garbage collected.
@changelog_1109_li
@changelog_1114_li
Foreign key constraints that refer to a quoted column did not work.
@changelog_1110_li
@changelog_1115_li
New meta data column INFORMATION_SCHEMA.TABLES.LAST_MODIFICATION to get the table modification counter.
@changelog_1111_li
@changelog_1116_li
Shell: line comments didn't work correctly.
@changelog_1112_li
@changelog_1117_li
H2 Console: Columns are now listed for up to 500 tables instead of 100.
@changelog_1113_li
@changelog_1118_li
H2 Console: Cmd+Enter executes the current statement, Alt+Space for autocomplete.
@changelog_1114_li
@changelog_1119_li
JaQu: the maximum length of a column can now be defined using maxLength. For an example, see Product.java (maxLength(category, 255)).
@changelog_1115_li
@changelog_1120_li
R&\#305;dvan A&\#287;ar has completed the Turkish translation of the H2 Console. Thanks a lot!
@changelog_1116_h2
@changelog_1121_h2
Version 1.1.104 (2008-11-28)
@changelog_1117_li
@changelog_1122_li
If a query that was used like a table contained group by and was ordered by an expression that is not in the column list, an exception was thrown.
@changelog_1118_li
@changelog_1123_li
JaQu: tables are now auto-created when running a query.
@changelog_1119_li
@changelog_1124_li
The optimizer had problems with function tables (for example CSVREAD and FTL_SEARCH). A new system property h2.estimatedFunctionTableRows (default 1000) defines how many rows can be expected in the table.
@changelog_1120_li
@changelog_1125_li
The function SUM could overflow when using large values. It returns now a data type that is safe.
@changelog_1121_li
@changelog_1126_li
The function AVG could overflow when using large values. Fixed.
@changelog_1122_li
@changelog_1127_li
The emergency reserve file has been removed. It didn't provide an appropriate solution for the problem. It is still possible for an application to detect and deal with the low disk space problem (deleting temporary files for example) using DatabaseEventListener.diskSpaceIsLow, but this method is now always called with stillAvailable=0.
@changelog_1123_li
@changelog_1128_li
Build: JAVA_HOME is now automatically detected on Mac OS X.
@changelog_1124_li
@changelog_1129_li
Testing for local connections was very slow on some systems.
@changelog_1125_li
@changelog_1130_li
The cache memory usage calculation is more conservative.
@changelog_1126_li
@changelog_1131_li
Allocating space got slower and slower the larger the database.
@changelog_1127_li
@changelog_1132_li
ALTER TABLE ALTER COLUMN could throw the wrong exception in the last version (Table not found).
@changelog_1128_li
@changelog_1133_li
Updatable result sets: the key columns can now be updated.
@changelog_1129_li
@changelog_1134_li
The H2DatabaseProvider for ActiveObjects is now included in the tools section.
@changelog_1130_li
@changelog_1135_li
The H2Platform for Oracle Toplink Essential has been improved a bit.
@changelog_1131_li
@changelog_1136_li
The Windows service to start H2 didn't work in version 1.1.
@changelog_1132_li
@changelog_1137_li
File systems with a maximum file size (for example FAT) are now supported using the file prefix 'split:'. In this case the files are split in parts of 1 GB. Example URL: jdbc:h2:split:~/db/test. If you want to split into parts of 1 MB, use jdbc:h2:split:20:~/db/test (the part size is 1 << x, the default is 30 meaning 1 GB).
@changelog_1133_li
@changelog_1138_li
The database now tries to detect if the classloader or virtual machine has almost shut down by checking if static final variables are set to null. This should help reduce exceptions when stopping the web application.
@changelog_1134_li
@changelog_1139_li
Compatibility for MS SQL Server DATEDIFF(YYYY, .., ..)
@changelog_1135_li
@changelog_1140_li
ResultSet.getObject for CLOB or BLOB will return a java.sql.Clob / java.sql.Blob object instead of a java.io.Reader / java.io.InputStream as in version 1.0. This behavior can be changed using the system property h2.returnLobObjects (true by default for version 1.1).
@changelog_1136_li
@changelog_1141_li
The interface CloseListener has a new method 'remove' that is called when the trigger is dropped.
@changelog_1137_li
@changelog_1142_li
Fulltext search: there was a memory leak when creating and dropping fulltext indexes in a loop.
@changelog_1138_h2
@changelog_1143_h2
Version 1.1.103 (2008-11-07)
@changelog_1139_li
@changelog_1144_li
Could not order by a formula when the formula was in the group by list but not in the select list.
@changelog_1140_li
@changelog_1145_li
Date values that match the daylight saving time end were not allowed in times zones were the daylight saving time ends at midnight, for years larger than 2037. Example: timezone Brasilia, date 2042-10-12. This is a problem of Java, however a workaround is implemented in H2 that solves most problems (except the problems of java.util.Date itself).
@changelog_1141_li
@changelog_1146_li
ALTER TABLE used a lot of memory when using multi-version concurrency.
@changelog_1142_li
@changelog_1147_li
Referential integrity for in-memory databases didn't work in some cases in version 1.1.102.
@changelog_1143_li
@changelog_1148_li
New column INFORMATION_SCHEMA.COLUMNS.SEQUENCE_NAME to get the name of the sequence for auto-increment columns.
@changelog_1144_li
@changelog_1149_li
Aliases for built-in data types (such as MEDIUMBLOB which is an alias for BLOB) can now be re-mapped to another data type using CREATE DOMAIN. However main built-in data types (such as INTEGER) can not be re-mapped.
@changelog_1145_li
@changelog_1150_li
The Japanese translation has been completed by Masahiro Ikemoto. Thanks a lot!
@changelog_1146_li
@changelog_1151_li
Improved PostgreSQL compatibility for NEXTVAL and CURRVAL.
@changelog_1147_li
@changelog_1152_li
Less heap memory is needed when multiple databases are open at the same time: The memory reserve (used to rollback after out of memory) is now global and no longer allocated for each database separately.
@changelog_1148_li
@changelog_1153_li
New system property h2.browser to set the browser to use.
@changelog_1149_li
@changelog_1154_li
To start the browser, java.awt.Desktop.browse is now used if available.
@changelog_1150_h2
@changelog_1155_h2
Version 1.1.102 (2008-10-24)
@changelog_1151_li
@changelog_1156_li
The French translation of the H2 Console has been improved by Olivier Parent. Thanks a lot!
@changelog_1152_li
@changelog_1157_li
There was a memory leak when creating and dropping tables and indexes in a loop (persistent database only).
@changelog_1153_li
@changelog_1158_li
SET LOG 2 was not effective if executed after opening the database.
@changelog_1154_li
@changelog_1159_li
Translating the H2 Console is now simpler.
@changelog_1155_li
@changelog_1160_li
Common exception (error code 23*) are no longer written to the .trace.db file by default.
@changelog_1156_li
@changelog_1161_li
In-memory databases don't write LOBs to files any longer.
@changelog_1157_li
@changelog_1162_li
Self referencing constraints didn't restrict deleting rows that reference itself if there is another row that references it.
@changelog_1158_li
@changelog_1163_li
ResultSetMetaData.getColumnName now returns the alias name except for columns.
@changelog_1159_li
@changelog_1164_li
Temporary files are now deleted when the database is closed, even if they were not garbage collected so far.
@changelog_1160_h2
@changelog_1165_h2
Version 1.1.101 (2008-10-17)
@changelog_1161_li
@changelog_1166_li
Errors with code 42000 - 42999 are no longer written to the trace file by default.
@changelog_1162_li
@changelog_1167_li
Queries with more than 10 tables are now faster.
@changelog_1163_li
@changelog_1168_li
Opening a connection with AUTO_SERVER=TRUE is now fast when the database is already open in another process (less than 0.01 seconds instead of 2 seconds).
@changelog_1164_li
@changelog_1169_li
IF [NOT] EXISTS is supported for named constraints in ALTER TABLE ... ADD/DROP CONSTRAINT.
@changelog_1165_li
@changelog_1170_li
The error messages have been translated to Spanish by Dario V. Fassi. Thanks a lot!
@changelog_1166_li
@changelog_1171_li
Linked tables: the automatic connection sharing didn't work. Actually the system property h2.shareLinkedConnections was working in the opposite direction: it was disabled when set to true. Now it works as expected.
@changelog_1167_li
@changelog_1172_li
Opening large database is now faster.
@changelog_1168_li
@changelog_1173_li
New system property h2.socketConnectTimeout, the timeout in milliseconds to connect to a server. The default is 2000 (2 seconds).
@changelog_1169_li
@changelog_1174_li
The wrong parameters were bound to subqueries with parameters, specially when using IN(SELECT ...) and IN(...).
@changelog_1170_li
@changelog_1175_li
Unset parameters were not detected when the query was re-compiled.
@changelog_1171_li
@changelog_1176_li
New functions ISO_YEAR, ISO_WEEK, ISO_DAY_OF_WEEK. Thanks a lot to Robert Rathsack for implementing those!
@changelog_1172_li
@changelog_1177_li
The date functions DAYOFYEAR, DAYOFMONTH, DAYOFWEEK are now called DAY_OF_YEAR, DAY_OF_MONTH, DAY_OF_WEEK (the old names still work).
@changelog_1173_li
@changelog_1178_li
An out of memory error while deleting or updating many rows could result in a strange exception.
@changelog_1174_li
@changelog_1179_li
Linked tables: compatibility with MS SQL Server has been improved.
@changelog_1175_li
@changelog_1180_li
Renaming tables that have foreign keys with cascade didn't work correctly.
@changelog_1176_li
@changelog_1181_li
The auto-reconnect feature didn't work when using the auto-server mode. Fixed.
@changelog_1177_li
@changelog_1182_li
Fulltext search: new method FT_DROP_INDEX.
@changelog_1178_li
@changelog_1183_li
The optimization to group using an index didn't work in some cases in version 1.1 (see also system property h2.optimizeGroupSorted).
@changelog_1179_li
@changelog_1184_li
OSGi meta data is included in the manifest file. An OSGi BundleActivator is included: it loads the database driver when starting the bundle, and unloads it when stopping the bundle.
@changelog_1180_li
@changelog_1185_li
The default value for MAX_MEMORY_UNDO is now 50000.
@changelog_1181_li
@changelog_1186_li
For alias columns, ResultSetMetaData.getTableName() and getColumnName() now return the real table and column name in the default mode.
@changelog_1182_li
@changelog_1187_li
In SQL scripts created with SCRIPT TO, schemas are now only created if they don't exist yet.
@changelog_1183_li
@changelog_1188_li
After re-connecting to a database, the database event listener (if set) is informed about it.
@changelog_1184_li
@changelog_1189_li
Local temporary tables now support indexes. Thanks a lot to Matt Roy!
@changelog_1185_li
@changelog_1190_li
RUNSCRIPT no longer uses a temporary file.
@changelog_1186_li
@changelog_1191_li
New system table INFORMATION_SCHEMA.SESSION_STATE containing the SQL statements that make up the session state. The list currently contains variables (SET @..) and local temporary tables (without data).
@changelog_1187_li
@changelog_1192_li
After an automatic re-connect, part of the session state stays (the part that is stored in the SESSION_STATE table).
@changelog_1188_li
@changelog_1193_li
The build didn't work if the directory temp didn't exist before.
@changelog_1189_li
@changelog_1194_li
New system property h2.maxReconnect (default 3) to limit the number of re-connects for the same SQL statement (this is usually only important for SHUTDOWN).
@changelog_1190_li
@changelog_1195_li
WHERE .. IN (SELECT ...) could throw a NullPointerException.
@changelog_1191_li
@changelog_1196_li
Improved Glassfish / Toplink support in H2Platform thanks to Marcio Borges from Brazil. Thanks a lot!
@changelog_1192_h2
@changelog_1197_h2
Version 1.1.100 (2008-10-04)
@changelog_1193_li
@changelog_1198_li
In version 1.1, the following system properties are now enabled by default: h2.lobFilesInDirectories, h2.optimizeGroupSorted, h2.optimizeInJoin, h2.shareLinkedConnections
@changelog_1194_li
@changelog_1199_li
The H2 Console tool now works with the JDBC-ODBC bridge.
@changelog_1195_li
@changelog_1200_li
The H2 Console tool now supports command line options to start things separately.
@changelog_1196_li
@changelog_1201_li
Large objects did not work for in-memory databases in server mode in Linux.
@changelog_1197_li
@changelog_1202_li
Connections from a local address other than 'localhost' were not allowed if remote connections were disabled. This was always a problem, but only got visible in the last release because the server no longer connects to 'localhost' if networked.
@changelog_1198_li
@changelog_1203_li
The h2console.war can now be built using the Java build.
@changelog_1199_li
@changelog_1204_li
By default, databases are shared in the same process. For read-only databases this causes unnecessary synchronization, but safes memory. If you want that each connection opens its own database, append ;OPEN_NEW=TRUE to the database URL.
@changelog_1200_li
@changelog_1205_li
New auto-reconnect feature will cause the JDBC driver to reconnect to the database if the connection is lost. To enable, append ;AUTO_RECONNECT=TRUE to the database URL. This is specially helpful when using AUTO_SERVER. AUTO_SERVER automatically uses auto-reconnect.
@changelog_1201_li
@changelog_1206_li
CreateCluster: the property 'serverlist' is now called 'serverList'.
@changelog_1202_li
@changelog_1207_li
The ConvertTraceFile tool could not parse some files because the trace mechanism did not encode prepared statement parameters.
@changelog_1203_li
@changelog_1208_li
Databases names can now be one character long (the minimum size used to be 2 characters).
@changelog_1204_h2
@changelog_1209_h2
Version 1.0.79 (2008-09-26)
@changelog_1205_li
@changelog_1210_li
Linked tables that point to the same database can now share the connection within the same database. Access to the same connection is serialized. To enable this feature, set the system property h2.shareLinkedConnections to true.
@changelog_1206_li
@changelog_1211_li
Multiple processes can now access the same database without having to explicitly start the server. To do that, append ;AUTO_SERVER=TRUE to the database URL. In this case, the server is started automatically if the connection is in embedded mode, and the server mode is used if a server is running. If the process that opened the first connection is closed, the other client need to reconnect (there is no automatic re-connect so far). Remote connections are allowed, but only to this database.
@changelog_1207_li
@changelog_1212_li
The server tool now displays the correct IP address if networked.
@changelog_1208_li
@changelog_1213_li
Can now start a TCP server with port 0 (automatically select a port).
@changelog_1209_li
@changelog_1214_li
Result sets with just a unique index can now be updated (previously a primary key was required).
@changelog_1210_li
@changelog_1215_li
LINKED TABLE: the schema name can now be set. When multiple tables exist in different schema, and the schema name is not set, an exception is thrown.
@changelog_1211_li
@changelog_1216_li
LINKED TABLE: worked around a bug in Oracle with the CHAR data type.
@changelog_1212_li
@changelog_1217_li
Faster hash code calculation for large binary arrays.
@changelog_1213_li
@changelog_1218_li
Faster storage re-use algorithm thanks to Greg Dhuse from cleversafe.com.
@changelog_1214_li
@changelog_1219_li
The database supports the SHOW command for better MySQL and PostgreSQL compatibility.
@changelog_1215_li
@changelog_1220_li
The H2 Console now abbreviates large texts in results.
@changelog_1216_li
@changelog_1221_li
Multiple UNION queries could not be used in derived tables.
@changelog_1217_li
@changelog_1222_li
Linked tables can now be read-only.
@changelog_1218_li
@changelog_1223_li
Temporary linked tables are now supported.
@changelog_1219_li
@changelog_1224_li
It was possible to create tables in read-only databases.
@changelog_1220_li
@changelog_1225_li
SET SCHEMA_SEARCH_PATH is now documented.
@changelog_1221_li
@changelog_1226_li
SET SCHEMA did not work for views.
@changelog_1222_li
@changelog_1227_li
Row level locking for MVCC is now enabled. The exception 'Concurrent update in table ...' is still thrown, but only after the lock timeout.
@changelog_1223_li
@changelog_1228_li
The maximum log file size setting was ignored for large databases.
@changelog_1224_li
@changelog_1229_li
Multi-Version Concurrency (MVCC) may no longer be used when using the multi-threaded kernel feature (MULTI_THREADED). An exception is thrown when trying to connect with both settings. Additional synchronization is required before those features can be used together.
@changelog_1225_li
@changelog_1230_li
The data type JAVA_OBJECT could not be used in updatable result sets.
@changelog_1226_li
@changelog_1231_li
The system property h2.optimizeInJoin did not work correctly.
@changelog_1227_li
@changelog_1232_li
Conditions such as ID=? AND ID>? were slow.
@changelog_1228_h2
@changelog_1233_h2
Version 1.0.78 (2008-08-28)
@changelog_1229_li
@changelog_1234_li
The documentation no longer uses a frameset (except the Javadocs).
@changelog_1230_li
@changelog_1235_li
When using DB_CLOSE_DELAY, sometimes a NullPointerException is thrown when the database is opened almost at the same time as it is closed automatically. Thanks a lot to Dmitry Pekar for finding this!
@changelog_1231_li
@changelog_1236_li
Java methods with variable number of parameters can now be used (for Java 1.5 or newer).
@changelog_1232_li
@changelog_1237_li
The Japanese translation has been improved by Masahiro Ikemoto. Thanks a lot!
@changelog_1233_li
@changelog_1238_li
The H2 Console replaced an empty user name with a single space.
@changelog_1234_li
@changelog_1239_li
The build target 'build jarSmall' now includes the embedded database.
@changelog_1235_li
@changelog_1240_li
JdbcDataSource now keeps the password in a char array where possible.
@changelog_1236_li
@changelog_1241_li
ResultSet.absolute did not always work with large result sets.
@changelog_1237_li
@changelog_1242_li
Column aliases can now be used in GROUP BY and HAVING.
@changelog_1238_li
@changelog_1243_li
Jason Brittain has contributed MySQL date functions. Thanks a lot! They are not in the h2.jar file currently, but in src/tools/org/h2/mode/FunctionsMySQL.java. To install, add this class to the classpath and call FunctionsMySQL.register(conn) in the Java code.
@changelog_1239_h2
@changelog_1244_h2
Version 1.0.77 (2008-08-16)
@changelog_1240_li
@changelog_1245_li
JaQu is now using prepared statements and supports Date, Time, Timestamp.
@changelog_1241_li
@changelog_1246_li
When using remote in-memory databases, large LOB objects did not work.
@changelog_1242_li
@changelog_1247_li
Timestamp columns such as TIMESTAMP(6) were not compatible to other database.
@changelog_1243_li
@changelog_1248_li
Opening a large database was slow if there was a problem opening the previous time.
@changelog_1244_li
@changelog_1249_li
NOT IN(SELECT ...) was incorrect if the subquery returns no rows.
@changelog_1245_li
@changelog_1250_li
CREATE TABLE AS SELECT did not work correctly in the multi-version concurrency mode.
@changelog_1246_li
@changelog_1251_li
Support a comma before closing a list, as in: create table test(id int,)
@changelog_1247_li
@changelog_1252_li
MySQL compatibility: linked tables had lower case column names on some systems.
@changelog_1248_li
@changelog_1253_li
DB2 compatibility: the DB2 fetch-first-clause is supported.
@changelog_1249_li
@changelog_1254_li
Oracle compatibility: old style outer join syntax using (+) did work correctly sometimes.
@changelog_1250_li
@changelog_1255_li
ResultSet.setFetchSize is now supported.
@changelog_1251_li
@changelog_1256_li
It has been reported that when using Install4j on some Linux systems and enabling the 'pack200' option, the h2.jar becomes corrupted by the install process, causing application failure. A workaround is to add an empty file h2.jar.nopack next to the h2.jar file. The reason for this problem is not known.
@changelog_1252_h2
@changelog_1257_h2
Version 1.0.76 (2008-07-27)
@changelog_1253_li
@changelog_1258_li
The comment of a domain (user defined data type) is now used as the default column comment when creating a column with this domain.
@changelog_1254_li
@changelog_1259_li
Invalid database names are now detected and a better error message is thrown.
@changelog_1255_li
@changelog_1260_li
ResultSetMetaData.getColumnClassName now returns the correct class name for BLOB and CLOB.
@changelog_1256_li
@changelog_1261_li
Fixed the Oracle mode: Oracle allows multiple rows only where all columns of the unique index are NULL.
@changelog_1257_li
@changelog_1262_li
There is a problem with Hibernate when using Boolean columns. A patch for Hibernate has been submitted at http://opensource.atlassian.com/projects/hibernate/browse/HHH-3401
@changelog_1258_li
@changelog_1263_li
ORDER BY on tableName.columnName didn't work correctly if the column name was also used as an alias.
@changelog_1259_li
@changelog_1264_li
H2 Console: The progress display when opening a database has been improved.
@changelog_1260_li
@changelog_1265_li
The error message when the server doesn't start has been improved.
@changelog_1261_li
@changelog_1266_li
Key values can now be changed in updatable result sets.
@changelog_1262_li
@changelog_1267_li
Changes in updatable result sets are now visible even when resetting the result set.
@changelog_1263_li
@changelog_1268_li
Temporary files were sometimes deleted too late when executing large insert, update, or delete operations.
@changelog_1264_li
@changelog_1269_li
The database file was growing after deleting many rows, and after large update operations.
@changelog_1265_h2
@changelog_1270_h2
Version 1.0.75 (2008-07-14)
@changelog_1266_li
@changelog_1271_li
Multi version concurrency (MVCC): when a row was updated or deleted, but this change was rolled back, the row was not visible by other sessions if no index was used to access it. Fixed.
@changelog_1267_li
@changelog_1272_li
Views with multiple joined tables (where one was an outer join) couldn't be used in some cases. Fixed.
@changelog_1268_li
@changelog_1273_li
The CSVREAD method did not process NULL correctly when using a whitespace field separator.
@changelog_1269_li
@changelog_1274_li
Fixed the Oracle mode: Oracle allows multiple rows with NULL in a unique index.
@changelog_1270_li
@changelog_1275_li
Running out of memory could result in incomplete transactions or corrupted databases. Fixed.
@changelog_1271_li
@changelog_1276_li
When using order by in a query that uses the same table multiple times, the order could be incorrect. Fixed.
@changelog_1272_li
@changelog_1277_li
Referential constraint checking improvement: now the constraint is only checked if the key column values change.
@changelog_1273_li
@changelog_1278_li
Some database metadata calls returned the wrong data type for DATA_TYPE columns.
@changelog_1274_li
@changelog_1279_li
The Lucene fulltext index was empty when opening a database with fulltext index enabled, and re-indexing it didn't work. Fixed.
@changelog_1275_li
@changelog_1280_li
The character '$' could not be used in identifier names (table name, column names and so on). Fixed.
@changelog_1276_li
@changelog_1281_li
The new method org.h2.tools.Server.startWebServer(conn) starts the H2 Console to inspect a database while debugging.
@changelog_1277_li
@changelog_1282_li
Stopping a WebServer didn't always work. Fixed.
@changelog_1278_h2
@changelog_1283_h2
Version 1.0.74 (2008-06-21)
@changelog_1279_li
@changelog_1284_li
Work on row level locking has been started (but there is nothing usable yet).
@changelog_1280_li
@changelog_1285_li
JaQu (Java Query), a tool similar to LINQ (Language Integrated Query; from Microsoft) is now included under src/tools/org/h2/jaqu. A small sample application is included under src/test/org/h2/test/jaqu.
@changelog_1281_li
@changelog_1286_li
The source code is now switched to Java 1.6 by default. To switch back to Java 1.4, run 'build compile'. The h2.jar file is still Java 1.4.
@changelog_1282_li
@changelog_1287_li
The ChangePassword tool is now called ChangeFileEncryption.
@changelog_1283_li
@changelog_1288_li
It is no longer allowed to create columns with the data type NULL. Also, it is no longer allowed to convert a column to the data type NULL. This was possible before but caused data loss.
@changelog_1284_li
@changelog_1289_li
When using computed columns or default values with a different data type than the column data type, a class cast exception could occur. Fixed.
@changelog_1285_li
@changelog_1290_li
Opening databases larger than 1 GB was sometimes very slow if a lot of data was deleted previously. Fixed.
@changelog_1286_li
@changelog_1291_li
RUNSCRIPT could throw a NullPointerException if the script name was an expression.
@changelog_1287_li
@changelog_1292_li
Improved compatibility. New compatibility modes for Oracle and Derby. New compatibility flag uniqueIndexNullDistinct to only allow one row with 'NULL' in a unique index. This flag is enabled for Derby, Oracle, MSSQLServer, and HSQLDB.
@changelog_1288_li
@changelog_1293_li
Linked tables: To view the statements that are executed against the target table, set the trace level to 3.
@changelog_1289_li
@changelog_1294_li
RunScript tool: new options to show and check the results of queries.
@changelog_1290_li
@changelog_1295_li
Deadlocks are now detected. One transaction is rolled back automatically.
@changelog_1291_li
@changelog_1296_li
The Lucene fulltext index was always re-created when opening a database with fulltext index enabled.
@changelog_1292_li
@changelog_1297_li
Support for overloaded Java methods. A user defined function can now be bound to multiple Java methods, if the Java methods have the same name but a different number of parameters. Thanks to Gary Tong for providing a patch!
@changelog_1293_h2
@changelog_1298_h2
Version 1.0.73 (2008-05-31)
@changelog_1294_li
@changelog_1299_li
ParameterMetaData now returns the right data type for most conditions, as in WHERE ID=?.
@changelog_1295_li
@changelog_1300_li
The table SYSTEM_RANGE now supports expressions and parameters.
@changelog_1296_li
@changelog_1301_li
New column INFORMATION_SCHEMA.CONSTRAINTS.UNIQUE_INDEX_NAME that contains the name of the unique index used to enforce this constraint, if there is such an index.
@changelog_1297_li
@changelog_1302_li
SET QUERY_TIMEOUT and Statement.setQueryTimeout no longer commits a transaction. The same applies to SET @VARIABLE, SET LOCK_TIMEOUT, SET TRACE_LEVEL_*, SET THROTTLE, and SET PATH.
@changelog_1298_li
@changelog_1303_li
The SCRIPT command does now emit IF NOT EXISTS for CREATE ROLE.
@changelog_1299_li
@changelog_1304_li
MySQL compatibility: auto_increment column are no longer automatically converted to primary key columns.
@changelog_1300_li
@changelog_1305_li
PostgreSQL compatibility: support for BOOL_OR and BOOL_AND aggregate functions.
@changelog_1301_li
@changelog_1306_li
Negative scale values for DECIMAL or NUMBER columns are now supported in regular tables and in linked tables.
@changelog_1302_li
@changelog_1307_li
A role or right can now be granted or revoked multiple times without getting an exception.
@changelog_1303_li
@changelog_1308_li
Infinite numbers in SQL scripts are listed as POWER(0, -1)), negative infinite as (-POWER(0, -1)), and NaN (not a number) as SQRT(-1).
@changelog_1304_li
@changelog_1309_li
The special double and float values 'NaN' (not a number) did not work correctly when sorting or comparing.
@changelog_1305_li
@changelog_1310_li
The fulltext search did not support CLOB data types.
@changelog_1306_li
@changelog_1311_li
If the drive with the database files was disconnected or unmounted while writing, sometimes a stack overflow exception was thrown instead of a IO exception.
@changelog_1307_li
@changelog_1312_li
The H2 Console could not be shut down from within the tool if the browser supports keepAlive (most browsers do).
@changelog_1308_li
@changelog_1313_li
If the password was passed as a char array, it was kept in an internal buffer longer than required. Theoretically the password could have been stolen if the main memory was swapped to disk before the garbage collection was run.
@changelog_1309_h2
@changelog_1314_h2
Version 1.0.72 (2008-05-10)
@changelog_1310_li
@changelog_1315_li
Some databases could not be opened when appending ;RECOVER=1 to the database URL.
@changelog_1311_li
@changelog_1316_li
The Japanese translation of the error messages and the H2 Console has been completed by Masahiro Ikemoto (Arizona Design Inc.)
@changelog_1312_li
@changelog_1317_li
Updates made to updatable rows are now visible within the same result set. DatabaseMetaData.ownUpdatesAreVisible now returns true.
@changelog_1313_li
@changelog_1318_li
ParameterMetaData now returns the correct data for INSERT and UPDATE statements.
@changelog_1314_li
@changelog_1319_li
H2 Shell: DESCRIBE now supports an schema name.
@changelog_1315_li
@changelog_1320_li
A subset of the PostgreSQL 'dollar quoting' feature is now supported.
@changelog_1316_li
@changelog_1321_li
SLF4J is now supported by using adding TRACE_LEVEL_FILE=4 to the database URL.
@changelog_1317_li
@changelog_1322_li
The recovery tool did not work if the table name contained spaces or if there was a comment on the table.
@changelog_1318_li
@changelog_1323_li
Triggers are no longer executed when changing the table structure (ALTER TABLE).
@changelog_1319_li
@changelog_1324_li
When setting BLOB or CLOB values larger than 65 KB using a remote connection, temporary files were kept on the client longer than required (until the connection was closed or the object is garbage collected). Now they are removed as soon as the PreparedStatement is closed, or when the value is overwritten.
@changelog_1320_li
@changelog_1325_li
Statements can now be canceled remotely (when using remote connections).
@changelog_1321_li
@changelog_1326_li
The Shell tool now uses java.io.Console to read the password when using JDK 1.6
@changelog_1322_li
@changelog_1327_li
When using read-only databases and setting LOG=2, an exception was written to the trace file when closing the database. Fixed.
@changelog_1323_h2
@changelog_1328_h2
Version 1.0.71 (2008-04-25)
@changelog_1324_li
@changelog_1329_li
H2 is now dual-licensed under the Eclipse Public License (EPL) and the old 'H2 License' (which is basically MPL).
@changelog_1325_li
@changelog_1330_li
Sometimes an exception 'File ID mismatch' or 'try to add a record twice' occurred after large records (8 KB or larger) are updated or deleted. See also http://code.google.com/p/h2database/issues/detail?id=22
@changelog_1326_li
@changelog_1331_li
H2 Console: The tools can now be translated (it didn't work in the last release).
@changelog_1327_li
@changelog_1332_li
New traditional Chinese translation. Thanks a lot to Derek Chao!
@changelog_1328_li
@changelog_1333_li
Indexes were not used when enabling the optimization for IN(SELECT...) (system property h2.optimizeInJoin).
@changelog_1329_h2
@changelog_1334_h2
Version 1.0.70 (2008-04-20)
@changelog_1330_li
@changelog_1335_li
The plan is to dual-license H2. The additional license is EPL (Eclipse Public License). The current license (MPL, Mozilla Public License) will stay. Current users are not affected because they can keep MPL. EPL is very similar to MPL, the only bigger difference is related to patents (EPL is a bit more business friendly in this regard). See also http://opensource.org/licenses/eclipse-1.0.php, http://www.eclipse.org/legal/eplfaq.php (FAQ), http://blogs.zdnet.com/Burnette/?p=131
@changelog_1331_li
@changelog_1336_li
Multi version concurrency (MVCC): when a row was updated, and the updated column was not indexed, this update was visible sometimes for other sessions even if it was not committed.
@changelog_1332_li
@changelog_1337_li
Calling SHUTDOWN on one connection and starting a query on another connection concurrently could result in a Java level deadlock.
@changelog_1333_li
@changelog_1338_li
New system property h2.enableAnonymousSSL (default: true) to enable anonymous SSL connections.
@changelog_1334_li
@changelog_1339_li
The precision if SUBSTR is now calculated if possible.
@changelog_1335_li
@changelog_1340_li
The autocomplete in the H2 Console has been improved a bit.
@changelog_1336_li
@changelog_1341_li
The tools in the H2 Console are now translatable.
@changelog_1337_li
@changelog_1342_li
The servlet and lucene jar files are now automatically downloaded when building.
@changelog_1338_li
@changelog_1343_li
The code switch tool has been replaced by a simpler tool called SwitchSource that just uses find and replace.
@changelog_1339_li
@changelog_1344_li
Started to write a Ant replacement ('JAnt') that uses pure Java build definitions. Advantages: ability to debug the build, extensible, flexible, no XML, a bit faster. Future plan: support creating custom h2 distributions (for embedded use). Maybe create a new project 'Jant' or 'Javen' if other people are interested.
@changelog_1340_li
@changelog_1345_li
The jar file is now about 10% smaller because the variable debugging info is no longer included. The source file and line number debugging info is still included. If required, the jar file size of the full version can be further reduced to about 720 KB using 'build jarSmall' or even more by removing unneeded components.
@changelog_1341_li
@changelog_1346_li
Added shell scripts run.sh and build.sh. chmod +x is required, but otherwise it should work. Feedback or improvements are welcome!
@changelog_1342_li
@changelog_1347_li
Databases in zip files: large queries are now supported. Temp files are created in the temp directory if required. The documentation how to create the zip file has been corrected.
The Japanese translation of the error messages and the H2 Console has been improved. Thanks a lot to Masahiro IKEMOTO.
@changelog_1345_li
@changelog_1350_li
Optimization for MIN() and MAX() when using MVCC.
@changelog_1346_li
@changelog_1351_li
To protect against remote brute force password attacks, the delay after each unsuccessful login now gets double as long. New system properties h2.delayWrongPasswordMin and h2.delayWrongPasswordMax.
@changelog_1347_li
@changelog_1352_li
After setting the query timeout and then resetting it, the next query would still timeout. Fixed.
@changelog_1348_li
@changelog_1353_li
Adding a IDENTITY column to a table with data threw a lock timeout.
@changelog_1349_li
@changelog_1354_li
OutOfMemoryError could occur when using EXISTS or IN(SELECT ..).
@changelog_1350_li
@changelog_1355_li
The built-in connection pool is not called JdbcConnectionPool. The API and documentation has been changed.
@changelog_1351_li
@changelog_1356_li
The ConvertTraceFile tool now generates SQL statement statistics at the end of the SQL script file (similar to the profiling data generated when using java -Xrunhprof).
@changelog_1352_li
@changelog_1357_li
Nested joins are now supported (A JOIN B JOIN C ON .. ON ..)
@download_1000_h1
...
...
@@ -4694,30 +4709,33 @@ Use PreparedStatements and cache them.
Related Projects
@jaqu_1031_a
Empire-db
Dreamsource ORM
@jaqu_1032_a
JEQUEL: Java Embedded QUEry Language
Empire-db
@jaqu_1033_a
Joist
JEQUEL: Java Embedded QUEry Language
@jaqu_1034_a
JoSQL
Joist
@jaqu_1035_a
LIQUidFORM
JoSQL
@jaqu_1036_a
Quaere (Alias implementation)
LIQUidFORM
@jaqu_1037_a
Quaere
Quaere (Alias implementation)
@jaqu_1038_a
Querydsl
Quaere
@jaqu_1039_a
Querydsl
@jaqu_1040_a
Squill
@license_1000_h1
...
...
@@ -5573,279 +5591,285 @@ Luntbuild
Build automation and management tool.
@links_1127_a
Magnolia
localdb
@links_1128_p
Microarray Data Management and Export System for PFGRC (Pathogen Functional Genomics Resource Center) Microarrays.
A tool that locates the full file path of the folder containing the database files.
@links_1129_a
MiniConnectionPoolManager
Magnolia
@links_1130_p
A lightweight standalone JDBC connection pool manager.
Microarray Data Management and Export System for PFGRC (Pathogen Functional Genomics Resource Center) Microarrays.
@links_1131_a
Mr. Persister
MiniConnectionPoolManager
@links_1132_p
Simple, small and fast object relational mapping.
A lightweight standalone JDBC connection pool manager.
@links_1133_a
Myna Application Server
Mr. Persister
@links_1134_p
Java web app that provides dynamic web content and Java libraries access from JavaScript.
Simple, small and fast object relational mapping.
@links_1135_a
MyTunesRss
Myna Application Server
@links_1136_p
MyTunesRSS lets you listen to your music wherever you are.
Java web app that provides dynamic web content and Java libraries access from JavaScript.
@links_1137_a
NCGC CurveFit
MyTunesRss
@links_1138_p
From: NIH Chemical Genomics Center, National Institutes of Health, USA. An open source application in the life sciences research field. This application handles chemical structures and biological responses of thousands of compounds with the potential to handle million+ compounds. It utilizes an embedded H2 database to enable flexible query/retrieval of all data including advanced chemical substructure and similarity searching. The application highlights an automated curve fitting and classification algorithm that outperforms commercial packages in the field. Commercial alternatives are typically small desktop software that handle a few dose response curves at a time. A couple of commercial packages that do handle several thousand curves are very expensive tools (>60k USD) that require manual curation of analysis by the user; require a license to Oracle; lack advanced query/retrieval; and the ability to handle chemical structures.
MyTunesRSS lets you listen to your music wherever you are.
@links_1139_a
Nuxeo
NCGC CurveFit
@links_1140_p
Standards-based, open source platform for building ECM applications.
From: NIH Chemical Genomics Center, National Institutes of Health, USA. An open source application in the life sciences research field. This application handles chemical structures and biological responses of thousands of compounds with the potential to handle million+ compounds. It utilizes an embedded H2 database to enable flexible query/retrieval of all data including advanced chemical substructure and similarity searching. The application highlights an automated curve fitting and classification algorithm that outperforms commercial packages in the field. Commercial alternatives are typically small desktop software that handle a few dose response curves at a time. A couple of commercial packages that do handle several thousand curves are very expensive tools (>60k USD) that require manual curation of analysis by the user; require a license to Oracle; lack advanced query/retrieval; and the ability to handle chemical structures.
@links_1141_a
nWire
Nuxeo
@links_1142_p
Eclipse plug-in which expedites Java development. It's main purpose is to help developers find code quicker and easily understand how it relates to the rest of the application, thus, understand the application structure.
Standards-based, open source platform for building ECM applications.
@links_1143_a
Ontology Works
nWire
@links_1144_p
This company provides semantic technologies including deductive information repositories (the Ontology Works Knowledge Servers), semantic information fusion and semantic federation of legacy databases, ontology-based domain modeling, and management of the distributed enterprise.
Eclipse plug-in which expedites Java development. It's main purpose is to help developers find code quicker and easily understand how it relates to the rest of the application, thus, understand the application structure.
@links_1145_a
Ontoprise OntoBroker
Ontology Works
@links_1146_p
SemanticWeb-Middleware. It supports all W3C Semantic Web recommendations: OWL, RDF, RDFS, SPARQL, and F-Logic.
This company provides semantic technologies including deductive information repositories (the Ontology Works Knowledge Servers), semantic information fusion and semantic federation of legacy databases, ontology-based domain modeling, and management of the distributed enterprise.
@links_1147_a
Open Anzo
Ontoprise OntoBroker
@links_1148_p
Semantic Application Server.
SemanticWeb-Middleware. It supports all W3C Semantic Web recommendations: OWL, RDF, RDFS, SPARQL, and F-Logic.
@links_1149_a
OpenGroove
Open Anzo
@links_1150_p
OpenGroove is a groupware program that allows users to synchronize data.
Semantic Application Server.
@links_1151_a
OpenSocial Development Environment (OSDE)
OpenGroove
@links_1152_p
Development tool for OpenSocial application.
OpenGroove is a groupware program that allows users to synchronize data.
@links_1153_a
Orion
OpenSocial Development Environment (OSDE)
@links_1154_p
J2EE Application Server.
Development tool for OpenSocial application.
@links_1155_a
P5H2
Orion
@links_1156_p
A library for the <a href="http://www.processing.org">Processing</a> programming language and environment.
J2EE Application Server.
@links_1157_a
Phase-6
P5H2
@links_1158_p
A computer based learning software.
A library for the <a href="http://www.processing.org">Processing</a> programming language and environment.
@links_1159_a
Pickle
Phase-6
@links_1160_p
Pickle is a Java library containing classes for persistence, concurrency, and logging.
A computer based learning software.
@links_1161_a
Piman
Pickle
@links_1162_p
Water treatment projects data management.
Pickle is a Java library containing classes for persistence, concurrency, and logging.
@links_1163_a
PolePosition
Piman
@links_1164_p
Open source database benchmark.
Water treatment projects data management.
@links_1165_a
Poormans
PolePosition
@links_1166_p
Very basic CMS running as a SWT application and generating static html pages.
Open source database benchmark.
@links_1167_a
Railo
Poormans
@links_1168_p
Railo is an alternative engine for the Cold Fusion Markup Language, that compiles code programmed in CFML into Java bytecode and executes it on a servlet engine.
Very basic CMS running as a SWT application and generating static html pages.
@links_1169_a
Razuna
Railo
@links_1170_p
Open source Digital Asset Management System with integrated Web Content Management.
Railo is an alternative engine for the Cold Fusion Markup Language, that compiles code programmed in CFML into Java bytecode and executes it on a servlet engine.
@links_1171_a
RIFE
Razuna
@links_1172_p
A full-stack web application framework with tools and APIs to implement most common web features.
Open source Digital Asset Management System with integrated Web Content Management.
@links_1173_a
Rutema
RIFE
@links_1174_p
Rutema is a test execution and management tool for heterogeneous development environments written in Ruby.
A full-stack web application framework with tools and APIs to implement most common web features.
@links_1175_a
Sava
Rutema
@links_1176_p
Open-source web-based content management system.
Rutema is a test execution and management tool for heterogeneous development environments written in Ruby.
@links_1177_a
Scriptella
Sava
@links_1178_p
ETL (Extract-Transform-Load) and script execution tool.
Open-source web-based content management system.
@links_1179_a
Sesar
Scriptella
@links_1180_p
Dependency Injection Container with Aspect Oriented Programming.
ETL (Extract-Transform-Load) and script execution tool.
@links_1181_a
SemmleCode
Sesar
@links_1182_p
Eclipse plugin to help you improve software quality.
Dependency Injection Container with Aspect Oriented Programming.
@links_1183_a
SeQuaLite
SemmleCode
@links_1184_p
A free, light-weight, java data access framework.
Eclipse plugin to help you improve software quality.
@links_1185_a
ShapeLogic
SeQuaLite
@links_1186_p
Toolkit for declarative programming, image processing and computer vision.
A free, light-weight, java data access framework.
@links_1187_a
Shellbook
ShapeLogic
@links_1188_p
Desktop publishing application.
Toolkit for declarative programming, image processing and computer vision.
@links_1189_a
Signsoft intelliBO
Shellbook
@links_1190_p
Persistence middleware supporting the JDO specification.
Desktop publishing application.
@links_1191_a
SimpleORM
Signsoft intelliBO
@links_1192_p
Simple Java Object Relational Mapping.
Persistence middleware supporting the JDO specification.
@links_1193_a
SymmetricDS
SimpleORM
@links_1194_p
A web-enabled, database independent, data synchronization/replication software.
Simple Java Object Relational Mapping.
@links_1195_a
SmartFoxServer
SymmetricDS
@links_1196_p
Platform for developing multiuser applications and games with Macromedia Flash.
A web-enabled, database independent, data synchronization/replication software.
@links_1197_a
Social Bookmarks Friend Finder
SmartFoxServer
@links_1198_p
A GUI application that allows you to find users with similar bookmarks to the user specified (for delicious.com).
Platform for developing multiuser applications and games with Macromedia Flash.
@links_1199_a
Springfuse
Social Bookmarks Friend Finder
@links_1200_p
Code generation For Spring, Spring MVC & Hibernate.
A GUI application that allows you to find users with similar bookmarks to the user specified (for delicious.com).
@links_1201_a
SQLOrm
Springfuse
@links_1202_p
Java Object Relation Mapping.
Code generation For Spring, Spring MVC & Hibernate.
@links_1203_a
StorYBook
SQLOrm
@links_1204_p
A summary-based tool for novelist and script writers. It helps to keep the overview over the various traces a story has.
Java Object Relation Mapping.
@links_1205_a
StreamCruncher
StorYBook
@links_1206_p
Event (stream) processing kernel.
A summary-based tool for novelist and script writers. It helps to keep the overview over the various traces a story has.
@links_1207_a
Tune Backup
StreamCruncher
@links_1208_p
Easy-to-use backup solution for your iTunes library.
Event (stream) processing kernel.
@links_1209_a
weblica
Tune Backup
@links_1210_p
Desktop CMS.
Easy-to-use backup solution for your iTunes library.
@links_1211_a
Web of Web
weblica
@links_1212_p
Collaborative and realtime interactive media platform for the web.
Desktop CMS.
@links_1213_a
Werkzeugkasten
Web of Web
@links_1214_p
Minimum Java Toolset.
Collaborative and realtime interactive media platform for the web.
@links_1215_a
VPDA
Werkzeugkasten
@links_1216_p
View providers driven applications is a Java based application framework for building applications composed from server components - view providers.
Minimum Java Toolset.
@links_1217_a
Volunteer database
VPDA
@links_1218_p
View providers driven applications is a Java based application framework for building applications composed from server components - view providers.
@links_1219_a
Volunteer database
@links_1220_p
A database front end to register volunteers, partnership and donation for a Non Profit organization.
@mainWeb_1000_h1
...
...
@@ -6152,7 +6176,7 @@ Performance Tuning
Performance Comparison
@performance_1007_p
In many cases H2 is a lot faster than other (open source and not open source) database engines. Please note this is mostly a single connection benchmark run on one computer.
In many cases H2 is faster than other (open source and not open source) database engines. Please note this is mostly a single connection benchmark run on one computer.
@performance_1008_h3
Embedded
...
...
@@ -6896,489 +6920,501 @@ The license of these databases does not allow to publish benchmark results. This
About this Benchmark
@performance_1255_h4
Number of Connections
How to Run
@performance_1256_p
This is mostly a single-connection benchmark. BenchB uses multiple connections; the other tests use one connection.
This test was executed as follows:
@performance_1257_h4
Real-World Tests
Separate Process per Database
@performance_1258_p
Good benchmarks emulate real-world use cases. This benchmark includes 3 test cases: A simple test case with one table and many small updates / deletes. BenchA is similar to the TPC-A test, but single connection / single threaded (see also: www.tpc.org). BenchB is similar to the TPC-B test, using multiple connections (one thread per connection). BenchC is similar to the TPC-C test, but single connection / single threaded.
For each database, a new process is started, to ensure the previous test does not impact the current test.
@performance_1259_h4
Comparing Embedded with Server Databases
Number of Connections
@performance_1260_p
This is mainly a benchmark for embedded databases (where the application runs in the same virtual machine as the database engine). However MySQL and PostgreSQL are not Java databases and cannot be embedded into a Java application. For the Java databases, both embedded and server modes are tested.
This is mostly a single-connection benchmark. BenchB uses multiple connections; the other tests use one connection.
@performance_1261_h4
Real-World Tests
@performance_1262_p
Good benchmarks emulate real-world use cases. This benchmark includes 3 test cases: A simple test case with one table and many small updates / deletes. BenchA is similar to the TPC-A test, but single connection / single threaded (see also: www.tpc.org). BenchB is similar to the TPC-B test, using multiple connections (one thread per connection). BenchC is similar to the TPC-C test, but single connection / single threaded.
@performance_1263_h4
Comparing Embedded with Server Databases
@performance_1264_p
This is mainly a benchmark for embedded databases (where the application runs in the same virtual machine as the database engine). However MySQL and PostgreSQL are not Java databases and cannot be embedded into a Java application. For the Java databases, both embedded and server modes are tested.
@performance_1265_h4
Test Platform
@performance_1262_p
@performance_1266_p
This test is run on Windows XP with the virus scanner switched off. The VM used is Sun JDK 1.5.
@performance_1263_h4
@performance_1267_h4
Multiple Runs
@performance_1264_p
@performance_1268_p
When a Java benchmark is run first, the code is not fully compiled and therefore runs slower than when running multiple times. A benchmark should always run the same test multiple times and ignore the first run(s). This benchmark runs three times, but only the last run is measured.
@performance_1265_h4
@performance_1269_h4
Memory Usage
@performance_1266_p
@performance_1270_p
It is not enough to measure the time taken, the memory usage is important as well. Performance can be improved by using a bigger cache, but the amount of memory is limited. HSQLDB tables are kept fully in memory by default; this benchmark uses 'disk based' tables for all databases. Unfortunately, it is not so easy to calculate the memory usage of PostgreSQL and MySQL, because they run in a different process than the test. This benchmark currently does not print memory usage of those databases.
@performance_1267_h4
@performance_1271_h4
Delayed Operations
@performance_1268_p
@performance_1272_p
Some databases delay some operations (for example flushing the buffers) until after the benchmark is run. This benchmark waits between each database tested, and each database runs in a different process (sequentially).
@performance_1269_h4
@performance_1273_h4
Transaction Commit / Durability
@performance_1270_p
@performance_1274_p
Durability means transaction committed to the database will not be lost. Some databases (for example MySQL) try to enforce this by default by calling fsync() to flush the buffers, but most hard drives don't actually flush all data. Calling fsync() slows down transaction commit a lot, but doesn't always make data durable. When comparing the results, it is important to think about the effect. Many database suggest to 'batch' operations when possible. This benchmark switches off autocommit when loading the data, and calls commit after each 1000 inserts. However many applications need 'short' transactions at runtime (a commit after each update). This benchmark commits after each update / delete in the simple benchmark, and after each business transaction in the other benchmarks. For databases that support delayed commits, a delay of one second is used.
@performance_1271_h4
@performance_1275_h4
Using Prepared Statements
@performance_1272_p
@performance_1276_p
Wherever possible, the test cases use prepared statements.
@performance_1273_h4
@performance_1277_h4
Currently Not Tested: Startup Time
@performance_1274_p
@performance_1278_p
The startup time of a database engine is important as well for embedded use. This time is not measured currently. Also, not tested is the time used to create a database and open an existing database. Here, one (wrapper) connection is opened at the start, and for each step a new connection is opened and then closed.
@performance_1275_h2
@performance_1279_h2
PolePosition Benchmark
@performance_1276_p
@performance_1280_p
The PolePosition is an open source benchmark. The algorithms are all quite simple. It was developed / sponsored by db4o.
@performance_1277_th
@performance_1281_th
Test Case
@performance_1278_th
@performance_1282_th
Unit
@performance_1279_th
@performance_1283_th
H2
@performance_1280_th
@performance_1284_th
HSQLDB
@performance_1281_th
@performance_1285_th
MySQL
@performance_1282_td
@performance_1286_td
Melbourne write
@performance_1283_td
@performance_1287_td
ms
@performance_1284_td
@performance_1288_td
369
@performance_1285_td
@performance_1289_td
249
@performance_1286_td
@performance_1290_td
2022
@performance_1287_td
@performance_1291_td
Melbourne read
@performance_1288_td
@performance_1292_td
ms
@performance_1289_td
@performance_1293_td
47
@performance_1290_td
@performance_1294_td
49
@performance_1291_td
@performance_1295_td
93
@performance_1292_td
@performance_1296_td
Melbourne read_hot
@performance_1293_td
@performance_1297_td
ms
@performance_1294_td
@performance_1298_td
24
@performance_1295_td
@performance_1299_td
43
@performance_1296_td
@performance_1300_td
95
@performance_1297_td
@performance_1301_td
Melbourne delete
@performance_1298_td
@performance_1302_td
ms
@performance_1299_td
@performance_1303_td
147
@performance_1300_td
@performance_1304_td
133
@performance_1301_td
@performance_1305_td
176
@performance_1302_td
@performance_1306_td
Sepang write
@performance_1303_td
@performance_1307_td
ms
@performance_1304_td
@performance_1308_td
965
@performance_1305_td
@performance_1309_td
1201
@performance_1306_td
@performance_1310_td
3213
@performance_1307_td
@performance_1311_td
Sepang read
@performance_1308_td
@performance_1312_td
ms
@performance_1309_td
@performance_1313_td
765
@performance_1310_td
@performance_1314_td
948
@performance_1311_td
@performance_1315_td
3455
@performance_1312_td
@performance_1316_td
Sepang read_hot
@performance_1313_td
@performance_1317_td
ms
@performance_1314_td
@performance_1318_td
789
@performance_1315_td
@performance_1319_td
859
@performance_1316_td
@performance_1320_td
3563
@performance_1317_td
@performance_1321_td
Sepang delete
@performance_1318_td
@performance_1322_td
ms
@performance_1319_td
@performance_1323_td
1384
@performance_1320_td
@performance_1324_td
1596
@performance_1321_td
@performance_1325_td
6214
@performance_1322_td
@performance_1326_td
Bahrain write
@performance_1323_td
@performance_1327_td
ms
@performance_1324_td
@performance_1328_td
1186
@performance_1325_td
@performance_1329_td
1387
@performance_1326_td
@performance_1330_td
6904
@performance_1327_td
@performance_1331_td
Bahrain query_indexed_string
@performance_1328_td
@performance_1332_td
ms
@performance_1329_td
@performance_1333_td
336
@performance_1330_td
@performance_1334_td
170
@performance_1331_td
@performance_1335_td
693
@performance_1332_td
@performance_1336_td
Bahrain query_string
@performance_1333_td
@performance_1337_td
ms
@performance_1334_td
@performance_1338_td
18064
@performance_1335_td
@performance_1339_td
39703
@performance_1336_td
@performance_1340_td
41243
@performance_1337_td
@performance_1341_td
Bahrain query_indexed_int
@performance_1338_td
@performance_1342_td
ms
@performance_1339_td
@performance_1343_td
104
@performance_1340_td
@performance_1344_td
134
@performance_1341_td
@performance_1345_td
678
@performance_1342_td
@performance_1346_td
Bahrain update
@performance_1343_td
@performance_1347_td
ms
@performance_1344_td
@performance_1348_td
191
@performance_1345_td
@performance_1349_td
87
@performance_1346_td
@performance_1350_td
159
@performance_1347_td
@performance_1351_td
Bahrain delete
@performance_1348_td
@performance_1352_td
ms
@performance_1349_td
@performance_1353_td
1215
@performance_1350_td
@performance_1354_td
729
@performance_1351_td
@performance_1355_td
6812
@performance_1352_td
@performance_1356_td
Imola retrieve
@performance_1353_td
@performance_1357_td
ms
@performance_1354_td
@performance_1358_td
198
@performance_1355_td
@performance_1359_td
194
@performance_1356_td
@performance_1360_td
4036
@performance_1357_td
@performance_1361_td
Barcelona write
@performance_1358_td
@performance_1362_td
ms
@performance_1359_td
@performance_1363_td
413
@performance_1360_td
@performance_1364_td
832
@performance_1361_td
@performance_1365_td
3191
@performance_1362_td
@performance_1366_td
Barcelona read
@performance_1363_td
@performance_1367_td
ms
@performance_1364_td
@performance_1368_td
119
@performance_1365_td
@performance_1369_td
160
@performance_1366_td
@performance_1370_td
1177
@performance_1367_td
@performance_1371_td
Barcelona query
@performance_1368_td
@performance_1372_td
ms
@performance_1369_td
@performance_1373_td
20
@performance_1370_td
@performance_1374_td
5169
@performance_1371_td
@performance_1375_td
101
@performance_1372_td
@performance_1376_td
Barcelona delete
@performance_1373_td
@performance_1377_td
ms
@performance_1374_td
@performance_1378_td
388
@performance_1375_td
@performance_1379_td
319
@performance_1376_td
@performance_1380_td
3287
@performance_1377_td
@performance_1381_td
Total
@performance_1378_td
@performance_1382_td
ms
@performance_1379_td
@performance_1383_td
26724
@performance_1380_td
@performance_1384_td
53962
@performance_1381_td
@performance_1385_td
87112
@performance_1382_p
@performance_1386_p
There are a few problems with the PolePosition test:
@performance_1383_li
@performance_1387_li
HSQLDB uses in-memory tables by default while H2 uses persistent tables. The HSQLDB version included in PolePosition does not support changing this, so you need to replace poleposition-0.20/lib/hsqldb.jar with a newer version (for example hsqldb-1.8.0.7.jar), and then use the setting hsqldb.connecturl=jdbc:hsqldb:file:data/hsqldb/dbbench2;hsqldb.default_table_type=cached;sql.enforce_size=true in Jdbc.properties.
@performance_1384_li
@performance_1388_li
HSQLDB keeps the database open between tests, while H2 closes the database (losing all the cache). To change that, use the database URL jdbc:h2:file:data/h2/dbbench;DB_CLOSE_DELAY=-1
@performance_1385_li
@performance_1389_li
The amount of cache memory is quite important, specially for the PolePosition test. Unfortunately, the PolePosition test does not take this into account.
@performance_1386_h2
@performance_1390_h2
Application Profiling
@performance_1387_h3
@performance_1391_h3
Analyze First
@performance_1388_p
@performance_1392_p
Before trying to optimize performance, it is important to understand where the problem is (what part of the application is slow). Blind optimization or optimization based on guesses should be avoided, because usually it is not an efficient strategy. There are various ways to analyze an application. Sometimes two implementations can be compared using System.currentTimeMillis(). But this does not work for complex applications with many modules, and for memory problems.
@performance_1389_p
@performance_1393_p
A good tool to measure both memory usage and performance is the <a href="http://www.yourkit.com">YourKit Java Profiler</a> .
@performance_1390_p
@performance_1394_p
A simple way to profile an application is to use the built-in profiling tool of java. Example:
@performance_1391_p
@performance_1395_p
Unfortunately, it is only possible to profile the application from start to end. Another solution is to create a number of full thread dumps. To do that, first run <code>jps -l</code> to get the process id, and then run <code>jstack <pid></code> or <code>kill -QUIT <pid></code> (Linux) or press Ctrl+C (Windows).
@performance_1392_h2
@performance_1396_h2
Database Profiling
@performance_1393_p
@performance_1397_p
The ConvertTraceFile tool generates SQL statement statistics at the end of the SQL script file. The format used is similar to the profiling data generated when using java -Xrunhprof. As an example, execute the the following script using the H2 Console:
@performance_1394_p
@performance_1398_p
Now convert the .trace.db file using the ConvertTraceFile tool:
@performance_1395_p
@performance_1399_p
The generated file <code>test.sql</code> will contain the SQL statements as well as the following profiling data (results vary):
@performance_1396_h2
@performance_1400_h2
Database Performance Tuning
@performance_1397_h3
@performance_1401_h3
Virus Scanners
@performance_1398_p
@performance_1402_p
Some virus scanners scan files every time they are accessed. It is very important for performance that database files are not scanned for viruses. The database engine does never interprets the data stored in the files as programs, that means even if somebody would store a virus in a database file, this would be harmless (when the virus does not run, it cannot spread). Some virus scanners allow to exclude files by suffix. Make sure files ending with .db are not scanned.
@performance_1399_h3
@performance_1403_h3
Using the Trace Options
@performance_1400_p
@performance_1404_p
If the main performance hot spots are in the database engine, in many cases the performance can be optimized by creating additional indexes, or changing the schema. Sometimes the application does not directly generate the SQL statements, for example if an O/R mapping tool is used. To view the SQL statements and JDBC API calls, you can use the trace options. For more information, see <a href="features.html#trace_options">Using the Trace Options</a> .
@performance_1401_h3
@performance_1405_h3
Index Usage
@performance_1402_p
@performance_1406_p
This database uses indexes to improve the performance of SELECT, UPDATE and DELETE statements. If a column is used in the WHERE clause of a query, and if an index exists on this column, then the index can be used. Multi-column indexes are used if all or the first columns of the index are used. Both equality lookup and range scans are supported. Indexes are used to order result sets, but only if the condition uses the same index or no index at all. The results are sorted in memory if required. Indexes are created automatically for primary key and unique constraints. Indexes are also created for foreign key constraints, if required. For other columns, indexes need to be created manually using the CREATE INDEX statement.
@performance_1403_h3
@performance_1407_h3
Optimizer
@performance_1404_p
@performance_1408_p
This database uses a cost based optimizer. For simple and queries and queries with medium complexity (less than 7 tables in the join), the expected cost (running time) of all possible plans is calculated, and the plan with the lowest cost is used. For more complex queries, the algorithm first tries all possible combinations for the first few tables, and the remaining tables added using a greedy algorithm (this works well for most joins). Afterwards a genetic algorithm is used to test at most 2000 distinct plans. Only left-deep plans are evaluated.
@performance_1405_h3
@performance_1409_h3
Expression Optimization
@performance_1406_p
@performance_1410_p
After the statement is parsed, all expressions are simplified automatically if possible. Operations are evaluated only once if all parameters are constant. Functions are also optimized, but only if the function is constant (always returns the same result for the same parameter values). If the WHERE clause is always false, then the table is not accessed at all.
@performance_1407_h3
@performance_1411_h3
COUNT(*) Optimization
@performance_1408_p
@performance_1412_p
If the query only counts all rows of a table, then the data is not accessed. However, this is only possible if no WHERE clause is used, that means it only works for queries of the form SELECT COUNT(*) FROM table.
When executing a query, at most one index per joined table can be used. If the same table is joined multiple times, for each join only one index is used. Example: for the query SELECT * FROM TEST T1, TEST T2 WHERE T1.NAME='A' AND T2.ID=T1.ID, two index can be used, in this case the index on NAME for T1 and the index on ID for T2.
@performance_1411_p
@performance_1415_p
If a table has multiple indexes, sometimes more than one index could be used. Example: if there is a table TEST(ID, NAME, FIRSTNAME) and an index on each column, then two indexes could be used for the query SELECT * FROM TEST WHERE NAME='A' AND FIRSTNAME='B', the index on NAME or the index on FIRSTNAME. It is not possible to use both indexes at the same time. Which index is used depends on the selectivity of the column. The selectivity describes the 'uniqueness' of values in a column. A selectivity of 100 means each value appears only once, and a selectivity of 1 means the same value appears in many or most rows. For the query above, the index on NAME should be used if the table contains more distinct names than first names.
@performance_1412_p
@performance_1416_p
The SQL statement ANALYZE can be used to automatically estimate the selectivity of the columns in the tables. This command should be run from time to time to improve the query plans generated by the optimizer.
@performance_1413_h3
@performance_1417_h3
Optimization Examples
@performance_1414_p
@performance_1418_p
See <code>src/test/org/h2/samples/optimizations.sql</code> for a few examples of queries that benefit from special optimizations built into the database.
@performance_1415_h3
@performance_1419_h3
Cache Size and Type
@performance_1416_p
@performance_1420_p
By default the cache size of H2 is quite small. Consider using a larger cache size, or enable the second level soft reference cache. See also <a href="features.html#cache_settings">Cache Settings</a> .
#H2 Console: command line properties are now longer stored in the properties file. They are now only used for the current process, except if they are explicitly saved.
#Identifiers with a digit and then a dollar sign now work. Example: A1$B.
@changelog_1003_li
#MS SQL Server compatibility: support for linked tables with NVARCHAR, NCHAR, NCLOB, and LONGNVARCHAR.
@changelog_1004_li
#Android: Workaround for a problem when using read-only databases in zip files (skip seems to be implemented incorrectly on the Android system).
@changelog_1005_li
#Calling execute() or prepareStatement() with null as the SQL statement now throws an exception.
@changelog_1006_li
#Benchmark: the number of executed statements was incorrect. The H2 database was loaded at the beginning of the test to collect results, now it is loaded at the very end. Thanks to Fred Toussi from HSQLDB for reporting those problems. However the changed do not affect the relative performance.
@changelog_1007_li
#H2 Console: command line settings are now longer stored in the properties file. They are now only used for the current process, except if they are explicitly saved.
@changelog_1008_li
#Cache: support for a second level soft-references cache. To enable it, append ;CACHE_TYPE=SOFT_LRU (or SOFT_TQ) to the database URL, or set the system property h2.cacheTypeDefault to "SOFT_LRU" / "SOFT_TQ". Enabling the second level cache reduces performance for small databases, but speeds up large databases. It makes sense to use it if the available memory size is unknown. Thanks a lot to Jan Kotek!
@changelog_1004_h2
@changelog_1009_h2
#Version 1.1.112 (2009-05-01)
@changelog_1005_li
@changelog_1010_li
#JdbcPreparedStatement.toString() could throw a NullPointerException.
#Connection pool: the default login timeout is now 5 minutes.
@changelog_1008_li
@changelog_1013_li
#After truncating tables, opening large databases could become slow because indexes were always re-built unnecessarily when opening.
@changelog_1009_li
@changelog_1014_li
#More bugs in the server-less multi-connection mode have been fixed: Sometimes parameters of prepared statements were lost when a reconnecting. Concurrent read operations were slow. To improve performance, executeQuery(..) must be used for queries (execute(..) switches to the write mode, which is slow).
@changelog_1010_li
@changelog_1015_li
#GROUP BY queries with a self-join (join to the same table) that were grouped by columns with indexes returned the wrong result in some cases.
@changelog_1011_li
@changelog_1016_li
#Improved error message when the .lock.db file modification time is in the future.
@changelog_1012_li
@changelog_1017_li
#The MERGE statement now returns 0 as the generated key if the row was updated.
@changelog_1013_li
@changelog_1018_li
#Running code coverage is now automated.
@changelog_1014_li
@changelog_1019_li
#A file system implementation can now be registered using FileSystem.register.
@changelog_1015_li
@changelog_1020_li
#The database file system is no longer included in the jar file, it moved to the test section.
@changelog_1016_h2
@changelog_1021_h2
#Version 1.1.111 (2009-04-10)
@changelog_1017_li
@changelog_1022_li
#In-memory databases can now run inside the Google App Engine.
@changelog_1018_li
@changelog_1023_li
#Queries that are ordered by an indexed column returned no rows in certain cases (if all rows were deleted from the table previously, and there is a low number of rows in the table, and when not using other conditions, and when using the default b tree index).
@changelog_1019_li
@changelog_1024_li
#The wrong exception was thrown when using unquoted text for the SQL statements COMMENT, ALTER USER, and SET PASSWORD.
@changelog_1020_li
@changelog_1025_li
#The built-in connection pool did not roll back transactions and enable autocommit enabled after closing a connection.
@changelog_1021_li
@changelog_1026_li
#Sometimes a StackOverflow occurred when checking for deadlock. See also http://code.google.com/p/h2database/issues/detail?id=61
@changelog_1022_li
@changelog_1027_li
#The Shell tool no longer truncates results with only one column, and displays a message if data was truncated.
@changelog_1023_h2
@changelog_1028_h2
#Version 1.1.110 (2009-04-03)
@changelog_1024_li
@changelog_1029_li
#Support for not persistent in-memory tables in regular (persistent) databases using CREATE MEMORY TABLE(..) NOT PERSISTENT. Thanks a lot to Sergi Vladykin for the patch!
@changelog_1025_li
@changelog_1030_li
#The H2 Console trimmed the password (removed leading and trailing spaces). This is no longer the case, to support encrypted H2 database with an empty user password.
@changelog_1026_li
@changelog_1031_li
#The data type of a SUBSTRING method was wrong.
@changelog_1027_li
@changelog_1032_li
#ResultSet.findColumn and get methods with column label parameters now also check for matching column names (like most databases except MySQL).
@changelog_1028_li
@changelog_1033_li
#H2 Console: the browser system property now supports a list of arguments. Example: java -Dh2.browser="open,-a,Safari,%url" ...
@changelog_1029_li
@changelog_1034_li
#Improved Javadoc navigation (similar to Scaladoc).
@changelog_1030_li
@changelog_1035_li
#H2 Console: auto-complete of identifiers did not work correctly for H2 databases in MySQL mode.
@changelog_1031_li
@changelog_1036_li
#DISTINCT and GROUP BY on a CLOB column was broken.
@changelog_1032_li
@changelog_1037_li
#The FTP server moved to the tools section and is no longer included in the h2*.jar file.
@changelog_1033_li
@changelog_1038_li
#Improved error message for unsupported features: now the message says what exactly is not supported.
@changelog_1034_li
@changelog_1039_li
#Improved OSGi support.
@changelog_1035_li
@changelog_1040_li
#Some internal caches did not use the LRU mechanism. Fixed (LOB file list, optimizer cost cache, trace system, view indexes, collection keys, compressed in-memory file system).
@changelog_1036_li
@changelog_1041_li
#The API of the tools changed a bit (each tool now returns an exit code).
@changelog_1037_li
@changelog_1042_li
#Command line help of the tools now match the javadocs. The build converts the javadocs to a resource that is read by the tool at runtime. This should not have an effect on using the database, but it reduces duplicate and out-of-sync documentation.
@changelog_1038_li
@changelog_1043_li
#CREATE TABLE: improved compatibility (support for UNIQUE NOT NULL).
@changelog_1039_li
@changelog_1044_li
#DatabaseMetaData.getSQLKeywords now returns the correct list.
@changelog_1040_li
@changelog_1045_li
#Deterministic user defined functions did not work when the parameter was a column. Fixed.
@changelog_1041_li
@changelog_1046_li
#JdbcConnectionPool.setLoginTimeout with 0 now uses the default timeout.
@changelog_1042_li
@changelog_1047_li
#Creating a JdbcConnectionPool has been simplified a bit.
@changelog_1043_li
@changelog_1048_li
#The built-in connection pool did not re-use connections. Getting a connection using the built-in JdbcConnectionPool is now about 70 times faster than opening connections using DriverManager.getConnection.
@changelog_1044_li
@changelog_1049_li
#More bugs in the server-less multi-connection mode have been fixed: If a process terminated while writing, other open connections were blocked. If two processes were writing to the database, sometimes the database was corrupt after closing.
@changelog_1045_li
@changelog_1050_li
#Linked tables to SQLite database can now be created.
@changelog_1046_li
@changelog_1051_li
#Nested IN(IN(...)) didn't work.
@changelog_1047_li
@changelog_1052_li
#NIO storage: the nio: prefix was using memory mapped files instead of FileChannel.
@changelog_1048_h2
@changelog_1053_h2
#Version 1.1.109 (2009-03-14)
@changelog_1049_li
@changelog_1054_li
#The optimization for IN(...) is now only used if comparing a column with an index.
@changelog_1050_li
@changelog_1055_li
#User defined functions can now be deterministic (see CREATE ALIAS documentation).
@changelog_1051_li
@changelog_1056_li
#Multiple nested queries in the FROM clause with parameters did not always work.
@changelog_1052_li
@changelog_1057_li
#When converting CLOB to BINARY, each character resulted in one byte. Now, the text is parsed as a hex as when converting VARCHAR.
@changelog_1053_li
@changelog_1058_li
#New experimental NIO storage mechanism with both FileChannel and memory mapped files. To use it, use the file name prefix nio: or nioMapped: as in jdbc:h2:nio:~/test. So far it looks like NIO storage is faster on Mac OS but slower on some Windows systems. Thanks a lot to Jan Kotek for the patch!
@changelog_1054_li
@changelog_1059_li
#The functions BITOR, BITAND, BITXOR, and MOD now accept and return BIGINT instead of INT.
@changelog_1055_li
@changelog_1060_li
#Could not use the same linked table multiple times in the same query.
@changelog_1056_li
@changelog_1061_li
#Bugs in the server-less multi-connection mode have been fixed.
@changelog_1057_li
@changelog_1062_li
#Column names could not be named "UNIQUE" (with the quotes).
@changelog_1058_li
@changelog_1063_li
#New system function TRANSACTION_ID() to get the current transaction identifier for a session.
@changelog_1059_h2
@changelog_1064_h2
#Version 1.1.108 (2009-02-28)
@changelog_1060_li
@changelog_1065_li
#When the shutdown hook closed the database, the last log file was deleted too early. This could cause uncommitted changes to be persisted. In some cases, this could cause data corruption.
@changelog_1061_li
@changelog_1066_li
#JdbcConnectionPool: it was possible to set a negative connection pool size.
@changelog_1062_li
@changelog_1067_li
#Fulltext search did not support table names with a backslash.
@changelog_1063_li
@changelog_1068_li
#The internal IntArray class did not work correctly when initialized with a zero length array.
@changelog_1064_li
@changelog_1069_li
#The H2 Console web application (war file) did only support ASCII characters. Now UTF-8 is supported.
@changelog_1065_li
@changelog_1070_li
#DATEADD does no longer require that the argument is a timestamp.
@changelog_1066_li
@changelog_1071_li
#The database file locking mechanism didn't work correctly on Mac OS.
@changelog_1067_li
@changelog_1072_li
#Some built-in functions reported the wrong precision, scale, and display size.
@changelog_1068_li
@changelog_1073_li
#MySQL compatibility for CREATE TABLE is improved (UNSIGNED, KEY).
@changelog_1069_li
@changelog_1074_li
#Recovery did not work if there were more than 255 lobs stored as files.
@changelog_1070_li
@changelog_1075_li
#New experimental mode to support multiple read-write connections without starting a server. To enable this mode, append ;FILE_LOCK=SERIALIZED;OPEN_NEW=TRUE to the database URL. Don't expect high performance when multiple concurrent writers.
@changelog_1071_li
@changelog_1076_li
#In a web application, the database classes are not unloaded if a connection is open. This may cause out of memory when re-deploying a web application. The DbStarter is changed to close all connections to the configured database (by executing SHUTDOWN).
@changelog_1072_li
@changelog_1077_li
#The WebServlet did not close the database when un-deploying the web application.
@changelog_1073_li
@changelog_1078_li
#The exception message of failed INSERT or MERGE statements now includes all values and the row number.
@changelog_1074_li
@changelog_1079_li
#If opening a database failed with an out of memory exception, some files were not closed.
@changelog_1075_li
@changelog_1080_li
#Optimizer: the expected runtime calculation was incorrect. The fixed calculation should give slightly better query plans when using many joins.
@changelog_1076_li
@changelog_1081_li
#Improved exception message when connecting to a just started server fails.
@changelog_1077_li
@changelog_1082_li
#Connection.isValid is a bit faster.
@changelog_1078_li
@changelog_1083_li
#H2 Console: The autocomplete feature has been improved a bit. It can now better parse conditions.
@changelog_1079_li
@changelog_1084_li
#When restarting a web application in Tomcat, an exception was thrown sometimes. In most cases this was a NullPointerException. A workaround in H2 has been implemented. The root cause of the problem is now documented in the FAQ: Tomcat sets all static fields (final or non-final) to null when unloading a web application. A workaround is to put the h2.jar in the lib directory, or set the system property org.apache.catalina.loader.WebappClassLoader.ENABLE_CLEAR_REFERENCES to false.
@changelog_1080_h2
@changelog_1085_h2
#Version 1.1.107 (2009-01-24)
@changelog_1081_li
@changelog_1086_li
#Some DatabaseMetaData operations did not work for non-admin users for versions 1.1.x.
@changelog_1082_li
@changelog_1087_li
#The MySQL compatibility extension fromUnixTime now used the English locale.
@changelog_1083_li
@changelog_1088_li
#When using LOG=2 and repeatedly updating the last row rows of a table, the index file grew quickly.
@changelog_1084_li
@changelog_1089_li
#In versions 1.1.105 and 1.1.106, encrypted script files of earlier versions could not be processed. This is now again possible. The problem was that such script files were stored in a special format (STORAGE=TEXT) but support for this format was removed in version 1.1.105.
@changelog_1085_li
@changelog_1090_li
#Enabling the trace mechanism by creating a specially named file is no longer supported.
@changelog_1086_h2
@changelog_1091_h2
#Version 1.1.106 (2009-01-04)
@changelog_1087_li
@changelog_1092_li
#Statement.setQueryTimeout did not work correctly for some statements.
@changelog_1088_li
@changelog_1093_li
#CREATE DOMAIN: Built-in data types can now only be changed if no tables exist.
@changelog_1089_li
@changelog_1094_li
#Linked tables: a workaround for Oracle DATE columns has been implemented.
@changelog_1090_li
@changelog_1095_li
#DatabaseMetaData.getPrimaryKeys: The column PK_NAME now contains the constraint name instead of the index name (compatibility for PostgreSQL and Derby).
@changelog_1091_li
@changelog_1096_li
#Using IN(..) inside a IN(SELECT..) did not always work.
@changelog_1092_li
@changelog_1097_li
#Views with IN(..) that used a view itself did not work.
@changelog_1093_li
@changelog_1098_li
#Union queries with LIMIT or ORDER BY that are used in a view or subquery did not work.
@changelog_1094_li
@changelog_1099_li
#The license change a bit: so far the license was modified to say 'Swiss law'. This is now changed back to the original 'US law'. This was requested by a user, and I don't see a problem.
@changelog_1095_li
@changelog_1100_li
#Constraints for local temporary tables now session scoped. So far they were global. Thanks a lot to Eric Faulhaber for finding and fixing this problem!
@changelog_1096_li
@changelog_1101_li
#When using the auto-server mode, and if the lock file was modified in the future, the wrong exception was thrown ('Connection is broken' instead of 'Error opening database: Lock file modified in the future').
@changelog_1097_h2
@changelog_1102_h2
#Version 1.1.105 (2008-12-19)
@changelog_1098_li
@changelog_1103_li
#The setting STORAGE=TEXT is no longer supported.
@changelog_1099_li
@changelog_1104_li
#Deleting a database using the tool DeleteDbFiles deleted LOB files of other databases in the same directory.
@changelog_1100_li
@changelog_1105_li
#When used in a subquery, LIKE and IN(..) did not work correctly sometimes.
@changelog_1101_li
@changelog_1106_li
#The fulltext search documentation has been improved.
@changelog_1102_li
@changelog_1107_li
#ARRAY_GET returned the wrong data type (ARRAY). Now it returns VARCHAR.
@changelog_1103_li
@changelog_1108_li
#Natural join: the joined columns are not repeated any more when using SELECT *.
@changelog_1104_li
@changelog_1109_li
#User defined aggregate functions: the method getType expected internal data types instead of SQL types.
@changelog_1105_li
@changelog_1110_li
#User defined aggregate functions did not work if there was no group by expression.
@changelog_1106_li
@changelog_1111_li
#MySQL compatibility: support for := assignment as in @sum:=@sum+x
@changelog_1107_li
@changelog_1112_li
#INSERT INTO TEST(SELECT * FROM TEST) is now supported.
@changelog_1108_li
@changelog_1113_li
#Each session threw an invisible exception when garbage collected.
@changelog_1109_li
@changelog_1114_li
#Foreign key constraints that refer to a quoted column did not work.
@changelog_1110_li
@changelog_1115_li
#New meta data column INFORMATION_SCHEMA.TABLES.LAST_MODIFICATION to get the table modification counter.
@changelog_1111_li
@changelog_1116_li
#Shell: line comments didn't work correctly.
@changelog_1112_li
@changelog_1117_li
#H2 Console: Columns are now listed for up to 500 tables instead of 100.
@changelog_1113_li
@changelog_1118_li
#H2 Console: Cmd+Enter executes the current statement, Alt+Space for autocomplete.
@changelog_1114_li
@changelog_1119_li
#JaQu: the maximum length of a column can now be defined using maxLength. For an example, see Product.java (maxLength(category, 255)).
@changelog_1115_li
@changelog_1120_li
#R&\#305;dvan A&\#287;ar has completed the Turkish translation of the H2 Console. Thanks a lot!
@changelog_1116_h2
@changelog_1121_h2
#Version 1.1.104 (2008-11-28)
@changelog_1117_li
@changelog_1122_li
#If a query that was used like a table contained group by and was ordered by an expression that is not in the column list, an exception was thrown.
@changelog_1118_li
@changelog_1123_li
#JaQu: tables are now auto-created when running a query.
@changelog_1119_li
@changelog_1124_li
#The optimizer had problems with function tables (for example CSVREAD and FTL_SEARCH). A new system property h2.estimatedFunctionTableRows (default 1000) defines how many rows can be expected in the table.
@changelog_1120_li
@changelog_1125_li
#The function SUM could overflow when using large values. It returns now a data type that is safe.
@changelog_1121_li
@changelog_1126_li
#The function AVG could overflow when using large values. Fixed.
@changelog_1122_li
@changelog_1127_li
#The emergency reserve file has been removed. It didn't provide an appropriate solution for the problem. It is still possible for an application to detect and deal with the low disk space problem (deleting temporary files for example) using DatabaseEventListener.diskSpaceIsLow, but this method is now always called with stillAvailable=0.
@changelog_1123_li
@changelog_1128_li
#Build: JAVA_HOME is now automatically detected on Mac OS X.
@changelog_1124_li
@changelog_1129_li
#Testing for local connections was very slow on some systems.
@changelog_1125_li
@changelog_1130_li
#The cache memory usage calculation is more conservative.
@changelog_1126_li
@changelog_1131_li
#Allocating space got slower and slower the larger the database.
@changelog_1127_li
@changelog_1132_li
#ALTER TABLE ALTER COLUMN could throw the wrong exception in the last version (Table not found).
@changelog_1128_li
@changelog_1133_li
#Updatable result sets: the key columns can now be updated.
@changelog_1129_li
@changelog_1134_li
#The H2DatabaseProvider for ActiveObjects is now included in the tools section.
@changelog_1130_li
@changelog_1135_li
#The H2Platform for Oracle Toplink Essential has been improved a bit.
@changelog_1131_li
@changelog_1136_li
#The Windows service to start H2 didn't work in version 1.1.
@changelog_1132_li
@changelog_1137_li
#File systems with a maximum file size (for example FAT) are now supported using the file prefix 'split:'. In this case the files are split in parts of 1 GB. Example URL: jdbc:h2:split:~/db/test. If you want to split into parts of 1 MB, use jdbc:h2:split:20:~/db/test (the part size is 1 << x, the default is 30 meaning 1 GB).
@changelog_1133_li
@changelog_1138_li
#The database now tries to detect if the classloader or virtual machine has almost shut down by checking if static final variables are set to null. This should help reduce exceptions when stopping the web application.
@changelog_1134_li
@changelog_1139_li
#Compatibility for MS SQL Server DATEDIFF(YYYY, .., ..)
@changelog_1135_li
@changelog_1140_li
#ResultSet.getObject for CLOB or BLOB will return a java.sql.Clob / java.sql.Blob object instead of a java.io.Reader / java.io.InputStream as in version 1.0. This behavior can be changed using the system property h2.returnLobObjects (true by default for version 1.1).
@changelog_1136_li
@changelog_1141_li
#The interface CloseListener has a new method 'remove' that is called when the trigger is dropped.
@changelog_1137_li
@changelog_1142_li
#Fulltext search: there was a memory leak when creating and dropping fulltext indexes in a loop.
@changelog_1138_h2
@changelog_1143_h2
#Version 1.1.103 (2008-11-07)
@changelog_1139_li
@changelog_1144_li
#Could not order by a formula when the formula was in the group by list but not in the select list.
@changelog_1140_li
@changelog_1145_li
#Date values that match the daylight saving time end were not allowed in times zones were the daylight saving time ends at midnight, for years larger than 2037. Example: timezone Brasilia, date 2042-10-12. This is a problem of Java, however a workaround is implemented in H2 that solves most problems (except the problems of java.util.Date itself).
@changelog_1141_li
@changelog_1146_li
#ALTER TABLE used a lot of memory when using multi-version concurrency.
@changelog_1142_li
@changelog_1147_li
#Referential integrity for in-memory databases didn't work in some cases in version 1.1.102.
@changelog_1143_li
@changelog_1148_li
#New column INFORMATION_SCHEMA.COLUMNS.SEQUENCE_NAME to get the name of the sequence for auto-increment columns.
@changelog_1144_li
@changelog_1149_li
#Aliases for built-in data types (such as MEDIUMBLOB which is an alias for BLOB) can now be re-mapped to another data type using CREATE DOMAIN. However main built-in data types (such as INTEGER) can not be re-mapped.
@changelog_1145_li
@changelog_1150_li
#The Japanese translation has been completed by Masahiro Ikemoto. Thanks a lot!
@changelog_1146_li
@changelog_1151_li
#Improved PostgreSQL compatibility for NEXTVAL and CURRVAL.
@changelog_1147_li
@changelog_1152_li
#Less heap memory is needed when multiple databases are open at the same time: The memory reserve (used to rollback after out of memory) is now global and no longer allocated for each database separately.
@changelog_1148_li
@changelog_1153_li
#New system property h2.browser to set the browser to use.
@changelog_1149_li
@changelog_1154_li
#To start the browser, java.awt.Desktop.browse is now used if available.
@changelog_1150_h2
@changelog_1155_h2
#Version 1.1.102 (2008-10-24)
@changelog_1151_li
@changelog_1156_li
#The French translation of the H2 Console has been improved by Olivier Parent. Thanks a lot!
@changelog_1152_li
@changelog_1157_li
#There was a memory leak when creating and dropping tables and indexes in a loop (persistent database only).
@changelog_1153_li
@changelog_1158_li
#SET LOG 2 was not effective if executed after opening the database.
@changelog_1154_li
@changelog_1159_li
#Translating the H2 Console is now simpler.
@changelog_1155_li
@changelog_1160_li
#Common exception (error code 23*) are no longer written to the .trace.db file by default.
@changelog_1156_li
@changelog_1161_li
#In-memory databases don't write LOBs to files any longer.
@changelog_1157_li
@changelog_1162_li
#Self referencing constraints didn't restrict deleting rows that reference itself if there is another row that references it.
@changelog_1158_li
@changelog_1163_li
#ResultSetMetaData.getColumnName now returns the alias name except for columns.
@changelog_1159_li
@changelog_1164_li
#Temporary files are now deleted when the database is closed, even if they were not garbage collected so far.
@changelog_1160_h2
@changelog_1165_h2
#Version 1.1.101 (2008-10-17)
@changelog_1161_li
@changelog_1166_li
#Errors with code 42000 - 42999 are no longer written to the trace file by default.
@changelog_1162_li
@changelog_1167_li
#Queries with more than 10 tables are now faster.
@changelog_1163_li
@changelog_1168_li
#Opening a connection with AUTO_SERVER=TRUE is now fast when the database is already open in another process (less than 0.01 seconds instead of 2 seconds).
@changelog_1164_li
@changelog_1169_li
#IF [NOT] EXISTS is supported for named constraints in ALTER TABLE ... ADD/DROP CONSTRAINT.
@changelog_1165_li
@changelog_1170_li
#The error messages have been translated to Spanish by Dario V. Fassi. Thanks a lot!
@changelog_1166_li
@changelog_1171_li
#Linked tables: the automatic connection sharing didn't work. Actually the system property h2.shareLinkedConnections was working in the opposite direction: it was disabled when set to true. Now it works as expected.
@changelog_1167_li
@changelog_1172_li
#Opening large database is now faster.
@changelog_1168_li
@changelog_1173_li
#New system property h2.socketConnectTimeout, the timeout in milliseconds to connect to a server. The default is 2000 (2 seconds).
@changelog_1169_li
@changelog_1174_li
#The wrong parameters were bound to subqueries with parameters, specially when using IN(SELECT ...) and IN(...).
@changelog_1170_li
@changelog_1175_li
#Unset parameters were not detected when the query was re-compiled.
@changelog_1171_li
@changelog_1176_li
#New functions ISO_YEAR, ISO_WEEK, ISO_DAY_OF_WEEK. Thanks a lot to Robert Rathsack for implementing those!
@changelog_1172_li
@changelog_1177_li
#The date functions DAYOFYEAR, DAYOFMONTH, DAYOFWEEK are now called DAY_OF_YEAR, DAY_OF_MONTH, DAY_OF_WEEK (the old names still work).
@changelog_1173_li
@changelog_1178_li
#An out of memory error while deleting or updating many rows could result in a strange exception.
@changelog_1174_li
@changelog_1179_li
#Linked tables: compatibility with MS SQL Server has been improved.
@changelog_1175_li
@changelog_1180_li
#Renaming tables that have foreign keys with cascade didn't work correctly.
@changelog_1176_li
@changelog_1181_li
#The auto-reconnect feature didn't work when using the auto-server mode. Fixed.
@changelog_1177_li
@changelog_1182_li
#Fulltext search: new method FT_DROP_INDEX.
@changelog_1178_li
@changelog_1183_li
#The optimization to group using an index didn't work in some cases in version 1.1 (see also system property h2.optimizeGroupSorted).
@changelog_1179_li
@changelog_1184_li
#OSGi meta data is included in the manifest file. An OSGi BundleActivator is included: it loads the database driver when starting the bundle, and unloads it when stopping the bundle.
@changelog_1180_li
@changelog_1185_li
#The default value for MAX_MEMORY_UNDO is now 50000.
@changelog_1181_li
@changelog_1186_li
#For alias columns, ResultSetMetaData.getTableName() and getColumnName() now return the real table and column name in the default mode.
@changelog_1182_li
@changelog_1187_li
#In SQL scripts created with SCRIPT TO, schemas are now only created if they don't exist yet.
@changelog_1183_li
@changelog_1188_li
#After re-connecting to a database, the database event listener (if set) is informed about it.
@changelog_1184_li
@changelog_1189_li
#Local temporary tables now support indexes. Thanks a lot to Matt Roy!
@changelog_1185_li
@changelog_1190_li
#RUNSCRIPT no longer uses a temporary file.
@changelog_1186_li
@changelog_1191_li
#New system table INFORMATION_SCHEMA.SESSION_STATE containing the SQL statements that make up the session state. The list currently contains variables (SET @..) and local temporary tables (without data).
@changelog_1187_li
@changelog_1192_li
#After an automatic re-connect, part of the session state stays (the part that is stored in the SESSION_STATE table).
@changelog_1188_li
@changelog_1193_li
#The build didn't work if the directory temp didn't exist before.
@changelog_1189_li
@changelog_1194_li
#New system property h2.maxReconnect (default 3) to limit the number of re-connects for the same SQL statement (this is usually only important for SHUTDOWN).
@changelog_1190_li
@changelog_1195_li
#WHERE .. IN (SELECT ...) could throw a NullPointerException.
@changelog_1191_li
@changelog_1196_li
#Improved Glassfish / Toplink support in H2Platform thanks to Marcio Borges from Brazil. Thanks a lot!
@changelog_1192_h2
@changelog_1197_h2
#Version 1.1.100 (2008-10-04)
@changelog_1193_li
@changelog_1198_li
#In version 1.1, the following system properties are now enabled by default: h2.lobFilesInDirectories, h2.optimizeGroupSorted, h2.optimizeInJoin, h2.shareLinkedConnections
@changelog_1194_li
@changelog_1199_li
#The H2 Console tool now works with the JDBC-ODBC bridge.
@changelog_1195_li
@changelog_1200_li
#The H2 Console tool now supports command line options to start things separately.
@changelog_1196_li
@changelog_1201_li
#Large objects did not work for in-memory databases in server mode in Linux.
@changelog_1197_li
@changelog_1202_li
#Connections from a local address other than 'localhost' were not allowed if remote connections were disabled. This was always a problem, but only got visible in the last release because the server no longer connects to 'localhost' if networked.
@changelog_1198_li
@changelog_1203_li
#The h2console.war can now be built using the Java build.
@changelog_1199_li
@changelog_1204_li
#By default, databases are shared in the same process. For read-only databases this causes unnecessary synchronization, but safes memory. If you want that each connection opens its own database, append ;OPEN_NEW=TRUE to the database URL.
@changelog_1200_li
@changelog_1205_li
#New auto-reconnect feature will cause the JDBC driver to reconnect to the database if the connection is lost. To enable, append ;AUTO_RECONNECT=TRUE to the database URL. This is specially helpful when using AUTO_SERVER. AUTO_SERVER automatically uses auto-reconnect.
@changelog_1201_li
@changelog_1206_li
#CreateCluster: the property 'serverlist' is now called 'serverList'.
@changelog_1202_li
@changelog_1207_li
#The ConvertTraceFile tool could not parse some files because the trace mechanism did not encode prepared statement parameters.
@changelog_1203_li
@changelog_1208_li
#Databases names can now be one character long (the minimum size used to be 2 characters).
@changelog_1204_h2
@changelog_1209_h2
#Version 1.0.79 (2008-09-26)
@changelog_1205_li
@changelog_1210_li
#Linked tables that point to the same database can now share the connection within the same database. Access to the same connection is serialized. To enable this feature, set the system property h2.shareLinkedConnections to true.
@changelog_1206_li
@changelog_1211_li
#Multiple processes can now access the same database without having to explicitly start the server. To do that, append ;AUTO_SERVER=TRUE to the database URL. In this case, the server is started automatically if the connection is in embedded mode, and the server mode is used if a server is running. If the process that opened the first connection is closed, the other client need to reconnect (there is no automatic re-connect so far). Remote connections are allowed, but only to this database.
@changelog_1207_li
@changelog_1212_li
#The server tool now displays the correct IP address if networked.
@changelog_1208_li
@changelog_1213_li
#Can now start a TCP server with port 0 (automatically select a port).
@changelog_1209_li
@changelog_1214_li
#Result sets with just a unique index can now be updated (previously a primary key was required).
@changelog_1210_li
@changelog_1215_li
#LINKED TABLE: the schema name can now be set. When multiple tables exist in different schema, and the schema name is not set, an exception is thrown.
@changelog_1211_li
@changelog_1216_li
#LINKED TABLE: worked around a bug in Oracle with the CHAR data type.
@changelog_1212_li
@changelog_1217_li
#Faster hash code calculation for large binary arrays.
@changelog_1213_li
@changelog_1218_li
#Faster storage re-use algorithm thanks to Greg Dhuse from cleversafe.com.
@changelog_1214_li
@changelog_1219_li
#The database supports the SHOW command for better MySQL and PostgreSQL compatibility.
@changelog_1215_li
@changelog_1220_li
#The H2 Console now abbreviates large texts in results.
@changelog_1216_li
@changelog_1221_li
#Multiple UNION queries could not be used in derived tables.
@changelog_1217_li
@changelog_1222_li
#Linked tables can now be read-only.
@changelog_1218_li
@changelog_1223_li
#Temporary linked tables are now supported.
@changelog_1219_li
@changelog_1224_li
#It was possible to create tables in read-only databases.
@changelog_1220_li
@changelog_1225_li
#SET SCHEMA_SEARCH_PATH is now documented.
@changelog_1221_li
@changelog_1226_li
#SET SCHEMA did not work for views.
@changelog_1222_li
@changelog_1227_li
#Row level locking for MVCC is now enabled. The exception 'Concurrent update in table ...' is still thrown, but only after the lock timeout.
@changelog_1223_li
@changelog_1228_li
#The maximum log file size setting was ignored for large databases.
@changelog_1224_li
@changelog_1229_li
#Multi-Version Concurrency (MVCC) may no longer be used when using the multi-threaded kernel feature (MULTI_THREADED). An exception is thrown when trying to connect with both settings. Additional synchronization is required before those features can be used together.
@changelog_1225_li
@changelog_1230_li
#The data type JAVA_OBJECT could not be used in updatable result sets.
@changelog_1226_li
@changelog_1231_li
#The system property h2.optimizeInJoin did not work correctly.
@changelog_1227_li
@changelog_1232_li
#Conditions such as ID=? AND ID>? were slow.
@changelog_1228_h2
@changelog_1233_h2
#Version 1.0.78 (2008-08-28)
@changelog_1229_li
@changelog_1234_li
#The documentation no longer uses a frameset (except the Javadocs).
@changelog_1230_li
@changelog_1235_li
#When using DB_CLOSE_DELAY, sometimes a NullPointerException is thrown when the database is opened almost at the same time as it is closed automatically. Thanks a lot to Dmitry Pekar for finding this!
@changelog_1231_li
@changelog_1236_li
#Java methods with variable number of parameters can now be used (for Java 1.5 or newer).
@changelog_1232_li
@changelog_1237_li
#The Japanese translation has been improved by Masahiro Ikemoto. Thanks a lot!
@changelog_1233_li
@changelog_1238_li
#The H2 Console replaced an empty user name with a single space.
@changelog_1234_li
@changelog_1239_li
#The build target 'build jarSmall' now includes the embedded database.
@changelog_1235_li
@changelog_1240_li
#JdbcDataSource now keeps the password in a char array where possible.
@changelog_1236_li
@changelog_1241_li
#ResultSet.absolute did not always work with large result sets.
@changelog_1237_li
@changelog_1242_li
#Column aliases can now be used in GROUP BY and HAVING.
@changelog_1238_li
@changelog_1243_li
#Jason Brittain has contributed MySQL date functions. Thanks a lot! They are not in the h2.jar file currently, but in src/tools/org/h2/mode/FunctionsMySQL.java. To install, add this class to the classpath and call FunctionsMySQL.register(conn) in the Java code.
@changelog_1239_h2
@changelog_1244_h2
#Version 1.0.77 (2008-08-16)
@changelog_1240_li
@changelog_1245_li
#JaQu is now using prepared statements and supports Date, Time, Timestamp.
@changelog_1241_li
@changelog_1246_li
#When using remote in-memory databases, large LOB objects did not work.
@changelog_1242_li
@changelog_1247_li
#Timestamp columns such as TIMESTAMP(6) were not compatible to other database.
@changelog_1243_li
@changelog_1248_li
#Opening a large database was slow if there was a problem opening the previous time.
@changelog_1244_li
@changelog_1249_li
#NOT IN(SELECT ...) was incorrect if the subquery returns no rows.
@changelog_1245_li
@changelog_1250_li
#CREATE TABLE AS SELECT did not work correctly in the multi-version concurrency mode.
@changelog_1246_li
@changelog_1251_li
#Support a comma before closing a list, as in: create table test(id int,)
@changelog_1247_li
@changelog_1252_li
#MySQL compatibility: linked tables had lower case column names on some systems.
@changelog_1248_li
@changelog_1253_li
#DB2 compatibility: the DB2 fetch-first-clause is supported.
@changelog_1249_li
@changelog_1254_li
#Oracle compatibility: old style outer join syntax using (+) did work correctly sometimes.
@changelog_1250_li
@changelog_1255_li
#ResultSet.setFetchSize is now supported.
@changelog_1251_li
@changelog_1256_li
#It has been reported that when using Install4j on some Linux systems and enabling the 'pack200' option, the h2.jar becomes corrupted by the install process, causing application failure. A workaround is to add an empty file h2.jar.nopack next to the h2.jar file. The reason for this problem is not known.
@changelog_1252_h2
@changelog_1257_h2
#Version 1.0.76 (2008-07-27)
@changelog_1253_li
@changelog_1258_li
#The comment of a domain (user defined data type) is now used as the default column comment when creating a column with this domain.
@changelog_1254_li
@changelog_1259_li
#Invalid database names are now detected and a better error message is thrown.
@changelog_1255_li
@changelog_1260_li
#ResultSetMetaData.getColumnClassName now returns the correct class name for BLOB and CLOB.
@changelog_1256_li
@changelog_1261_li
#Fixed the Oracle mode: Oracle allows multiple rows only where all columns of the unique index are NULL.
@changelog_1257_li
@changelog_1262_li
#There is a problem with Hibernate when using Boolean columns. A patch for Hibernate has been submitted at http://opensource.atlassian.com/projects/hibernate/browse/HHH-3401
@changelog_1258_li
@changelog_1263_li
#ORDER BY on tableName.columnName didn't work correctly if the column name was also used as an alias.
@changelog_1259_li
@changelog_1264_li
#H2 Console: The progress display when opening a database has been improved.
@changelog_1260_li
@changelog_1265_li
#The error message when the server doesn't start has been improved.
@changelog_1261_li
@changelog_1266_li
#Key values can now be changed in updatable result sets.
@changelog_1262_li
@changelog_1267_li
#Changes in updatable result sets are now visible even when resetting the result set.
@changelog_1263_li
@changelog_1268_li
#Temporary files were sometimes deleted too late when executing large insert, update, or delete operations.
@changelog_1264_li
@changelog_1269_li
#The database file was growing after deleting many rows, and after large update operations.
@changelog_1265_h2
@changelog_1270_h2
#Version 1.0.75 (2008-07-14)
@changelog_1266_li
@changelog_1271_li
#Multi version concurrency (MVCC): when a row was updated or deleted, but this change was rolled back, the row was not visible by other sessions if no index was used to access it. Fixed.
@changelog_1267_li
@changelog_1272_li
#Views with multiple joined tables (where one was an outer join) couldn't be used in some cases. Fixed.
@changelog_1268_li
@changelog_1273_li
#The CSVREAD method did not process NULL correctly when using a whitespace field separator.
@changelog_1269_li
@changelog_1274_li
#Fixed the Oracle mode: Oracle allows multiple rows with NULL in a unique index.
@changelog_1270_li
@changelog_1275_li
#Running out of memory could result in incomplete transactions or corrupted databases. Fixed.
@changelog_1271_li
@changelog_1276_li
#When using order by in a query that uses the same table multiple times, the order could be incorrect. Fixed.
@changelog_1272_li
@changelog_1277_li
#Referential constraint checking improvement: now the constraint is only checked if the key column values change.
@changelog_1273_li
@changelog_1278_li
#Some database metadata calls returned the wrong data type for DATA_TYPE columns.
@changelog_1274_li
@changelog_1279_li
#The Lucene fulltext index was empty when opening a database with fulltext index enabled, and re-indexing it didn't work. Fixed.
@changelog_1275_li
@changelog_1280_li
#The character '$' could not be used in identifier names (table name, column names and so on). Fixed.
@changelog_1276_li
@changelog_1281_li
#The new method org.h2.tools.Server.startWebServer(conn) starts the H2 Console to inspect a database while debugging.
@changelog_1277_li
@changelog_1282_li
#Stopping a WebServer didn't always work. Fixed.
@changelog_1278_h2
@changelog_1283_h2
#Version 1.0.74 (2008-06-21)
@changelog_1279_li
@changelog_1284_li
#Work on row level locking has been started (but there is nothing usable yet).
@changelog_1280_li
@changelog_1285_li
#JaQu (Java Query), a tool similar to LINQ (Language Integrated Query; from Microsoft) is now included under src/tools/org/h2/jaqu. A small sample application is included under src/test/org/h2/test/jaqu.
@changelog_1281_li
@changelog_1286_li
#The source code is now switched to Java 1.6 by default. To switch back to Java 1.4, run 'build compile'. The h2.jar file is still Java 1.4.
@changelog_1282_li
@changelog_1287_li
#The ChangePassword tool is now called ChangeFileEncryption.
@changelog_1283_li
@changelog_1288_li
#It is no longer allowed to create columns with the data type NULL. Also, it is no longer allowed to convert a column to the data type NULL. This was possible before but caused data loss.
@changelog_1284_li
@changelog_1289_li
#When using computed columns or default values with a different data type than the column data type, a class cast exception could occur. Fixed.
@changelog_1285_li
@changelog_1290_li
#Opening databases larger than 1 GB was sometimes very slow if a lot of data was deleted previously. Fixed.
@changelog_1286_li
@changelog_1291_li
#RUNSCRIPT could throw a NullPointerException if the script name was an expression.
@changelog_1287_li
@changelog_1292_li
#Improved compatibility. New compatibility modes for Oracle and Derby. New compatibility flag uniqueIndexNullDistinct to only allow one row with 'NULL' in a unique index. This flag is enabled for Derby, Oracle, MSSQLServer, and HSQLDB.
@changelog_1288_li
@changelog_1293_li
#Linked tables: To view the statements that are executed against the target table, set the trace level to 3.
@changelog_1289_li
@changelog_1294_li
#RunScript tool: new options to show and check the results of queries.
@changelog_1290_li
@changelog_1295_li
#Deadlocks are now detected. One transaction is rolled back automatically.
@changelog_1291_li
@changelog_1296_li
#The Lucene fulltext index was always re-created when opening a database with fulltext index enabled.
@changelog_1292_li
@changelog_1297_li
#Support for overloaded Java methods. A user defined function can now be bound to multiple Java methods, if the Java methods have the same name but a different number of parameters. Thanks to Gary Tong for providing a patch!
@changelog_1293_h2
@changelog_1298_h2
#Version 1.0.73 (2008-05-31)
@changelog_1294_li
@changelog_1299_li
#ParameterMetaData now returns the right data type for most conditions, as in WHERE ID=?.
@changelog_1295_li
@changelog_1300_li
#The table SYSTEM_RANGE now supports expressions and parameters.
@changelog_1296_li
@changelog_1301_li
#New column INFORMATION_SCHEMA.CONSTRAINTS.UNIQUE_INDEX_NAME that contains the name of the unique index used to enforce this constraint, if there is such an index.
@changelog_1297_li
@changelog_1302_li
#SET QUERY_TIMEOUT and Statement.setQueryTimeout no longer commits a transaction. The same applies to SET @VARIABLE, SET LOCK_TIMEOUT, SET TRACE_LEVEL_*, SET THROTTLE, and SET PATH.
@changelog_1298_li
@changelog_1303_li
#The SCRIPT command does now emit IF NOT EXISTS for CREATE ROLE.
@changelog_1299_li
@changelog_1304_li
#MySQL compatibility: auto_increment column are no longer automatically converted to primary key columns.
@changelog_1300_li
@changelog_1305_li
#PostgreSQL compatibility: support for BOOL_OR and BOOL_AND aggregate functions.
@changelog_1301_li
@changelog_1306_li
#Negative scale values for DECIMAL or NUMBER columns are now supported in regular tables and in linked tables.
@changelog_1302_li
@changelog_1307_li
#A role or right can now be granted or revoked multiple times without getting an exception.
@changelog_1303_li
@changelog_1308_li
#Infinite numbers in SQL scripts are listed as POWER(0, -1)), negative infinite as (-POWER(0, -1)), and NaN (not a number) as SQRT(-1).
@changelog_1304_li
@changelog_1309_li
#The special double and float values 'NaN' (not a number) did not work correctly when sorting or comparing.
@changelog_1305_li
@changelog_1310_li
#The fulltext search did not support CLOB data types.
@changelog_1306_li
@changelog_1311_li
#If the drive with the database files was disconnected or unmounted while writing, sometimes a stack overflow exception was thrown instead of a IO exception.
@changelog_1307_li
@changelog_1312_li
#The H2 Console could not be shut down from within the tool if the browser supports keepAlive (most browsers do).
@changelog_1308_li
@changelog_1313_li
#If the password was passed as a char array, it was kept in an internal buffer longer than required. Theoretically the password could have been stolen if the main memory was swapped to disk before the garbage collection was run.
@changelog_1309_h2
@changelog_1314_h2
#Version 1.0.72 (2008-05-10)
@changelog_1310_li
@changelog_1315_li
#Some databases could not be opened when appending ;RECOVER=1 to the database URL.
@changelog_1311_li
@changelog_1316_li
#The Japanese translation of the error messages and the H2 Console has been completed by Masahiro Ikemoto (Arizona Design Inc.)
@changelog_1312_li
@changelog_1317_li
#Updates made to updatable rows are now visible within the same result set. DatabaseMetaData.ownUpdatesAreVisible now returns true.
@changelog_1313_li
@changelog_1318_li
#ParameterMetaData now returns the correct data for INSERT and UPDATE statements.
@changelog_1314_li
@changelog_1319_li
#H2 Shell: DESCRIBE now supports an schema name.
@changelog_1315_li
@changelog_1320_li
#A subset of the PostgreSQL 'dollar quoting' feature is now supported.
@changelog_1316_li
@changelog_1321_li
#SLF4J is now supported by using adding TRACE_LEVEL_FILE=4 to the database URL.
@changelog_1317_li
@changelog_1322_li
#The recovery tool did not work if the table name contained spaces or if there was a comment on the table.
@changelog_1318_li
@changelog_1323_li
#Triggers are no longer executed when changing the table structure (ALTER TABLE).
@changelog_1319_li
@changelog_1324_li
#When setting BLOB or CLOB values larger than 65 KB using a remote connection, temporary files were kept on the client longer than required (until the connection was closed or the object is garbage collected). Now they are removed as soon as the PreparedStatement is closed, or when the value is overwritten.
@changelog_1320_li
@changelog_1325_li
#Statements can now be canceled remotely (when using remote connections).
@changelog_1321_li
@changelog_1326_li
#The Shell tool now uses java.io.Console to read the password when using JDK 1.6
@changelog_1322_li
@changelog_1327_li
#When using read-only databases and setting LOG=2, an exception was written to the trace file when closing the database. Fixed.
@changelog_1323_h2
@changelog_1328_h2
#Version 1.0.71 (2008-04-25)
@changelog_1324_li
@changelog_1329_li
#H2 is now dual-licensed under the Eclipse Public License (EPL) and the old 'H2 License' (which is basically MPL).
@changelog_1325_li
@changelog_1330_li
#Sometimes an exception 'File ID mismatch' or 'try to add a record twice' occurred after large records (8 KB or larger) are updated or deleted. See also http://code.google.com/p/h2database/issues/detail?id=22
@changelog_1326_li
@changelog_1331_li
#H2 Console: The tools can now be translated (it didn't work in the last release).
@changelog_1327_li
@changelog_1332_li
#New traditional Chinese translation. Thanks a lot to Derek Chao!
@changelog_1328_li
@changelog_1333_li
#Indexes were not used when enabling the optimization for IN(SELECT...) (system property h2.optimizeInJoin).
@changelog_1329_h2
@changelog_1334_h2
#Version 1.0.70 (2008-04-20)
@changelog_1330_li
@changelog_1335_li
#The plan is to dual-license H2. The additional license is EPL (Eclipse Public License). The current license (MPL, Mozilla Public License) will stay. Current users are not affected because they can keep MPL. EPL is very similar to MPL, the only bigger difference is related to patents (EPL is a bit more business friendly in this regard). See also http://opensource.org/licenses/eclipse-1.0.php, http://www.eclipse.org/legal/eplfaq.php (FAQ), http://blogs.zdnet.com/Burnette/?p=131
@changelog_1331_li
@changelog_1336_li
#Multi version concurrency (MVCC): when a row was updated, and the updated column was not indexed, this update was visible sometimes for other sessions even if it was not committed.
@changelog_1332_li
@changelog_1337_li
#Calling SHUTDOWN on one connection and starting a query on another connection concurrently could result in a Java level deadlock.
@changelog_1333_li
@changelog_1338_li
#New system property h2.enableAnonymousSSL (default: true) to enable anonymous SSL connections.
@changelog_1334_li
@changelog_1339_li
#The precision if SUBSTR is now calculated if possible.
@changelog_1335_li
@changelog_1340_li
#The autocomplete in the H2 Console has been improved a bit.
@changelog_1336_li
@changelog_1341_li
#The tools in the H2 Console are now translatable.
@changelog_1337_li
@changelog_1342_li
#The servlet and lucene jar files are now automatically downloaded when building.
@changelog_1338_li
@changelog_1343_li
#The code switch tool has been replaced by a simpler tool called SwitchSource that just uses find and replace.
@changelog_1339_li
@changelog_1344_li
#Started to write a Ant replacement ('JAnt') that uses pure Java build definitions. Advantages: ability to debug the build, extensible, flexible, no XML, a bit faster. Future plan: support creating custom h2 distributions (for embedded use). Maybe create a new project 'Jant' or 'Javen' if other people are interested.
@changelog_1340_li
@changelog_1345_li
#The jar file is now about 10% smaller because the variable debugging info is no longer included. The source file and line number debugging info is still included. If required, the jar file size of the full version can be further reduced to about 720 KB using 'build jarSmall' or even more by removing unneeded components.
@changelog_1341_li
@changelog_1346_li
#Added shell scripts run.sh and build.sh. chmod +x is required, but otherwise it should work. Feedback or improvements are welcome!
@changelog_1342_li
@changelog_1347_li
#Databases in zip files: large queries are now supported. Temp files are created in the temp directory if required. The documentation how to create the zip file has been corrected.
#The Japanese translation of the error messages and the H2 Console has been improved. Thanks a lot to Masahiro IKEMOTO.
@changelog_1345_li
@changelog_1350_li
#Optimization for MIN() and MAX() when using MVCC.
@changelog_1346_li
@changelog_1351_li
#To protect against remote brute force password attacks, the delay after each unsuccessful login now gets double as long. New system properties h2.delayWrongPasswordMin and h2.delayWrongPasswordMax.
@changelog_1347_li
@changelog_1352_li
#After setting the query timeout and then resetting it, the next query would still timeout. Fixed.
@changelog_1348_li
@changelog_1353_li
#Adding a IDENTITY column to a table with data threw a lock timeout.
@changelog_1349_li
@changelog_1354_li
#OutOfMemoryError could occur when using EXISTS or IN(SELECT ..).
@changelog_1350_li
@changelog_1355_li
#The built-in connection pool is not called JdbcConnectionPool. The API and documentation has been changed.
@changelog_1351_li
@changelog_1356_li
#The ConvertTraceFile tool now generates SQL statement statistics at the end of the SQL script file (similar to the profiling data generated when using java -Xrunhprof).
@changelog_1352_li
@changelog_1357_li
#Nested joins are now supported (A JOIN B JOIN C ON .. ON ..)
@download_1000_h1
...
...
@@ -4696,30 +4711,33 @@ Sourceファイル
#Related Projects
@jaqu_1031_a
#Empire-db
#Dreamsource ORM
@jaqu_1032_a
#JEQUEL: Java Embedded QUEry Language
#Empire-db
@jaqu_1033_a
#Joist
#JEQUEL: Java Embedded QUEry Language
@jaqu_1034_a
#JoSQL
#Joist
@jaqu_1035_a
#LIQUidFORM
#JoSQL
@jaqu_1036_a
#Quaere (Alias implementation)
#LIQUidFORM
@jaqu_1037_a
#Quaere
#Quaere (Alias implementation)
@jaqu_1038_a
#Querydsl
#Quaere
@jaqu_1039_a
#Querydsl
@jaqu_1040_a
#Squill
@license_1000_h1
...
...
@@ -5575,279 +5593,285 @@ Sourceファイル
#Build automation and management tool.
@links_1127_a
#Magnolia
#localdb
@links_1128_p
#Microarray Data Management and Export System for PFGRC (Pathogen Functional Genomics Resource Center) Microarrays.
#A tool that locates the full file path of the folder containing the database files.
@links_1129_a
#MiniConnectionPoolManager
#Magnolia
@links_1130_p
#A lightweight standalone JDBC connection pool manager.
#Microarray Data Management and Export System for PFGRC (Pathogen Functional Genomics Resource Center) Microarrays.
@links_1131_a
#Mr. Persister
#MiniConnectionPoolManager
@links_1132_p
#Simple, small and fast object relational mapping.
#A lightweight standalone JDBC connection pool manager.
@links_1133_a
#Myna Application Server
#Mr. Persister
@links_1134_p
#Java web app that provides dynamic web content and Java libraries access from JavaScript.
#Simple, small and fast object relational mapping.
@links_1135_a
#MyTunesRss
#Myna Application Server
@links_1136_p
#MyTunesRSS lets you listen to your music wherever you are.
#Java web app that provides dynamic web content and Java libraries access from JavaScript.
@links_1137_a
#NCGC CurveFit
#MyTunesRss
@links_1138_p
#From: NIH Chemical Genomics Center, National Institutes of Health, USA. An open source application in the life sciences research field. This application handles chemical structures and biological responses of thousands of compounds with the potential to handle million+ compounds. It utilizes an embedded H2 database to enable flexible query/retrieval of all data including advanced chemical substructure and similarity searching. The application highlights an automated curve fitting and classification algorithm that outperforms commercial packages in the field. Commercial alternatives are typically small desktop software that handle a few dose response curves at a time. A couple of commercial packages that do handle several thousand curves are very expensive tools (>60k USD) that require manual curation of analysis by the user; require a license to Oracle; lack advanced query/retrieval; and the ability to handle chemical structures.
#MyTunesRSS lets you listen to your music wherever you are.
@links_1139_a
#Nuxeo
#NCGC CurveFit
@links_1140_p
#Standards-based, open source platform for building ECM applications.
#From: NIH Chemical Genomics Center, National Institutes of Health, USA. An open source application in the life sciences research field. This application handles chemical structures and biological responses of thousands of compounds with the potential to handle million+ compounds. It utilizes an embedded H2 database to enable flexible query/retrieval of all data including advanced chemical substructure and similarity searching. The application highlights an automated curve fitting and classification algorithm that outperforms commercial packages in the field. Commercial alternatives are typically small desktop software that handle a few dose response curves at a time. A couple of commercial packages that do handle several thousand curves are very expensive tools (>60k USD) that require manual curation of analysis by the user; require a license to Oracle; lack advanced query/retrieval; and the ability to handle chemical structures.
@links_1141_a
#nWire
#Nuxeo
@links_1142_p
#Eclipse plug-in which expedites Java development. It's main purpose is to help developers find code quicker and easily understand how it relates to the rest of the application, thus, understand the application structure.
#Standards-based, open source platform for building ECM applications.
@links_1143_a
#Ontology Works
#nWire
@links_1144_p
#This company provides semantic technologies including deductive information repositories (the Ontology Works Knowledge Servers), semantic information fusion and semantic federation of legacy databases, ontology-based domain modeling, and management of the distributed enterprise.
#Eclipse plug-in which expedites Java development. It's main purpose is to help developers find code quicker and easily understand how it relates to the rest of the application, thus, understand the application structure.
@links_1145_a
#Ontoprise OntoBroker
#Ontology Works
@links_1146_p
#SemanticWeb-Middleware. It supports all W3C Semantic Web recommendations: OWL, RDF, RDFS, SPARQL, and F-Logic.
#This company provides semantic technologies including deductive information repositories (the Ontology Works Knowledge Servers), semantic information fusion and semantic federation of legacy databases, ontology-based domain modeling, and management of the distributed enterprise.
@links_1147_a
#Open Anzo
#Ontoprise OntoBroker
@links_1148_p
#Semantic Application Server.
#SemanticWeb-Middleware. It supports all W3C Semantic Web recommendations: OWL, RDF, RDFS, SPARQL, and F-Logic.
@links_1149_a
#OpenGroove
#Open Anzo
@links_1150_p
#OpenGroove is a groupware program that allows users to synchronize data.
#Semantic Application Server.
@links_1151_a
#OpenSocial Development Environment (OSDE)
#OpenGroove
@links_1152_p
#Development tool for OpenSocial application.
#OpenGroove is a groupware program that allows users to synchronize data.
@links_1153_a
#Orion
#OpenSocial Development Environment (OSDE)
@links_1154_p
#J2EE Application Server.
#Development tool for OpenSocial application.
@links_1155_a
#P5H2
#Orion
@links_1156_p
#A library for the <a href="http://www.processing.org">Processing</a> programming language and environment.
#J2EE Application Server.
@links_1157_a
#Phase-6
#P5H2
@links_1158_p
#A computer based learning software.
#A library for the <a href="http://www.processing.org">Processing</a> programming language and environment.
@links_1159_a
#Pickle
#Phase-6
@links_1160_p
#Pickle is a Java library containing classes for persistence, concurrency, and logging.
#A computer based learning software.
@links_1161_a
#Piman
#Pickle
@links_1162_p
#Water treatment projects data management.
#Pickle is a Java library containing classes for persistence, concurrency, and logging.
@links_1163_a
#PolePosition
#Piman
@links_1164_p
#Open source database benchmark.
#Water treatment projects data management.
@links_1165_a
#Poormans
#PolePosition
@links_1166_p
#Very basic CMS running as a SWT application and generating static html pages.
#Open source database benchmark.
@links_1167_a
#Railo
#Poormans
@links_1168_p
#Railo is an alternative engine for the Cold Fusion Markup Language, that compiles code programmed in CFML into Java bytecode and executes it on a servlet engine.
#Very basic CMS running as a SWT application and generating static html pages.
@links_1169_a
#Razuna
#Railo
@links_1170_p
#Open source Digital Asset Management System with integrated Web Content Management.
#Railo is an alternative engine for the Cold Fusion Markup Language, that compiles code programmed in CFML into Java bytecode and executes it on a servlet engine.
@links_1171_a
#RIFE
#Razuna
@links_1172_p
#A full-stack web application framework with tools and APIs to implement most common web features.
#Open source Digital Asset Management System with integrated Web Content Management.
@links_1173_a
#Rutema
#RIFE
@links_1174_p
#Rutema is a test execution and management tool for heterogeneous development environments written in Ruby.
#A full-stack web application framework with tools and APIs to implement most common web features.
@links_1175_a
#Sava
#Rutema
@links_1176_p
#Open-source web-based content management system.
#Rutema is a test execution and management tool for heterogeneous development environments written in Ruby.
@links_1177_a
#Scriptella
#Sava
@links_1178_p
#ETL (Extract-Transform-Load) and script execution tool.
#Open-source web-based content management system.
@links_1179_a
#Sesar
#Scriptella
@links_1180_p
#Dependency Injection Container with Aspect Oriented Programming.
#ETL (Extract-Transform-Load) and script execution tool.
@links_1181_a
#SemmleCode
#Sesar
@links_1182_p
#Eclipse plugin to help you improve software quality.
#Dependency Injection Container with Aspect Oriented Programming.
@links_1183_a
#SeQuaLite
#SemmleCode
@links_1184_p
#A free, light-weight, java data access framework.
#Eclipse plugin to help you improve software quality.
@links_1185_a
#ShapeLogic
#SeQuaLite
@links_1186_p
#Toolkit for declarative programming, image processing and computer vision.
#A free, light-weight, java data access framework.
@links_1187_a
#Shellbook
#ShapeLogic
@links_1188_p
#Desktop publishing application.
#Toolkit for declarative programming, image processing and computer vision.
@links_1189_a
#Signsoft intelliBO
#Shellbook
@links_1190_p
#Persistence middleware supporting the JDO specification.
#Desktop publishing application.
@links_1191_a
#SimpleORM
#Signsoft intelliBO
@links_1192_p
#Simple Java Object Relational Mapping.
#Persistence middleware supporting the JDO specification.
@links_1193_a
#SymmetricDS
#SimpleORM
@links_1194_p
#A web-enabled, database independent, data synchronization/replication software.
#Simple Java Object Relational Mapping.
@links_1195_a
#SmartFoxServer
#SymmetricDS
@links_1196_p
#Platform for developing multiuser applications and games with Macromedia Flash.
#A web-enabled, database independent, data synchronization/replication software.
@links_1197_a
#Social Bookmarks Friend Finder
#SmartFoxServer
@links_1198_p
#A GUI application that allows you to find users with similar bookmarks to the user specified (for delicious.com).
#Platform for developing multiuser applications and games with Macromedia Flash.
@links_1199_a
#Springfuse
#Social Bookmarks Friend Finder
@links_1200_p
#Code generation For Spring, Spring MVC & Hibernate.
#A GUI application that allows you to find users with similar bookmarks to the user specified (for delicious.com).
@links_1201_a
#SQLOrm
#Springfuse
@links_1202_p
#Java Object Relation Mapping.
#Code generation For Spring, Spring MVC & Hibernate.
@links_1203_a
#StorYBook
#SQLOrm
@links_1204_p
#A summary-based tool for novelist and script writers. It helps to keep the overview over the various traces a story has.
#Java Object Relation Mapping.
@links_1205_a
#StreamCruncher
#StorYBook
@links_1206_p
#Event (stream) processing kernel.
#A summary-based tool for novelist and script writers. It helps to keep the overview over the various traces a story has.
@links_1207_a
#Tune Backup
#StreamCruncher
@links_1208_p
#Easy-to-use backup solution for your iTunes library.
#Event (stream) processing kernel.
@links_1209_a
#weblica
#Tune Backup
@links_1210_p
#Desktop CMS.
#Easy-to-use backup solution for your iTunes library.
@links_1211_a
#Web of Web
#weblica
@links_1212_p
#Collaborative and realtime interactive media platform for the web.
#Desktop CMS.
@links_1213_a
#Werkzeugkasten
#Web of Web
@links_1214_p
#Minimum Java Toolset.
#Collaborative and realtime interactive media platform for the web.
@links_1215_a
#VPDA
#Werkzeugkasten
@links_1216_p
#View providers driven applications is a Java based application framework for building applications composed from server components - view providers.
#Minimum Java Toolset.
@links_1217_a
#Volunteer database
#VPDA
@links_1218_p
#View providers driven applications is a Java based application framework for building applications composed from server components - view providers.
@links_1219_a
#Volunteer database
@links_1220_p
#A database front end to register volunteers, partnership and donation for a Non Profit organization.
@mainWeb_1000_h1
...
...
@@ -6157,7 +6181,7 @@ See what this database can do and how to use these features.
#Performance Comparison
@performance_1007_p
#In many cases H2 is a lot faster than other (open source and not open source) database engines. Please note this is mostly a single connection benchmark run on one computer.
#In many cases H2 is faster than other (open source and not open source) database engines. Please note this is mostly a single connection benchmark run on one computer.
@performance_1008_h3
#Embedded
...
...
@@ -6901,489 +6925,501 @@ MySQL
#About this Benchmark
@performance_1255_h4
#Number of Connections
#How to Run
@performance_1256_p
#This is mostly a single-connection benchmark. BenchB uses multiple connections; the other tests use one connection.
#This test was executed as follows:
@performance_1257_h4
#Real-World Tests
#Separate Process per Database
@performance_1258_p
#Good benchmarks emulate real-world use cases. This benchmark includes 3 test cases: A simple test case with one table and many small updates / deletes. BenchA is similar to the TPC-A test, but single connection / single threaded (see also: www.tpc.org). BenchB is similar to the TPC-B test, using multiple connections (one thread per connection). BenchC is similar to the TPC-C test, but single connection / single threaded.
#For each database, a new process is started, to ensure the previous test does not impact the current test.
@performance_1259_h4
#Comparing Embedded with Server Databases
#Number of Connections
@performance_1260_p
#This is mainly a benchmark for embedded databases (where the application runs in the same virtual machine as the database engine). However MySQL and PostgreSQL are not Java databases and cannot be embedded into a Java application. For the Java databases, both embedded and server modes are tested.
#This is mostly a single-connection benchmark. BenchB uses multiple connections; the other tests use one connection.
@performance_1261_h4
#Test Platform
#Real-World Tests
@performance_1262_p
#Good benchmarks emulate real-world use cases. This benchmark includes 3 test cases: A simple test case with one table and many small updates / deletes. BenchA is similar to the TPC-A test, but single connection / single threaded (see also: www.tpc.org). BenchB is similar to the TPC-B test, using multiple connections (one thread per connection). BenchC is similar to the TPC-C test, but single connection / single threaded.
@performance_1263_h4
#Comparing Embedded with Server Databases
@performance_1264_p
#This is mainly a benchmark for embedded databases (where the application runs in the same virtual machine as the database engine). However MySQL and PostgreSQL are not Java databases and cannot be embedded into a Java application. For the Java databases, both embedded and server modes are tested.
@performance_1265_h4
#Test Platform
@performance_1266_p
#This test is run on Windows XP with the virus scanner switched off. The VM used is Sun JDK 1.5.
@performance_1263_h4
@performance_1267_h4
#Multiple Runs
@performance_1264_p
@performance_1268_p
#When a Java benchmark is run first, the code is not fully compiled and therefore runs slower than when running multiple times. A benchmark should always run the same test multiple times and ignore the first run(s). This benchmark runs three times, but only the last run is measured.
@performance_1265_h4
@performance_1269_h4
#Memory Usage
@performance_1266_p
@performance_1270_p
#It is not enough to measure the time taken, the memory usage is important as well. Performance can be improved by using a bigger cache, but the amount of memory is limited. HSQLDB tables are kept fully in memory by default; this benchmark uses 'disk based' tables for all databases. Unfortunately, it is not so easy to calculate the memory usage of PostgreSQL and MySQL, because they run in a different process than the test. This benchmark currently does not print memory usage of those databases.
@performance_1267_h4
@performance_1271_h4
#Delayed Operations
@performance_1268_p
@performance_1272_p
#Some databases delay some operations (for example flushing the buffers) until after the benchmark is run. This benchmark waits between each database tested, and each database runs in a different process (sequentially).
@performance_1269_h4
@performance_1273_h4
#Transaction Commit / Durability
@performance_1270_p
@performance_1274_p
#Durability means transaction committed to the database will not be lost. Some databases (for example MySQL) try to enforce this by default by calling fsync() to flush the buffers, but most hard drives don't actually flush all data. Calling fsync() slows down transaction commit a lot, but doesn't always make data durable. When comparing the results, it is important to think about the effect. Many database suggest to 'batch' operations when possible. This benchmark switches off autocommit when loading the data, and calls commit after each 1000 inserts. However many applications need 'short' transactions at runtime (a commit after each update). This benchmark commits after each update / delete in the simple benchmark, and after each business transaction in the other benchmarks. For databases that support delayed commits, a delay of one second is used.
@performance_1271_h4
@performance_1275_h4
#Using Prepared Statements
@performance_1272_p
@performance_1276_p
#Wherever possible, the test cases use prepared statements.
@performance_1273_h4
@performance_1277_h4
#Currently Not Tested: Startup Time
@performance_1274_p
@performance_1278_p
#The startup time of a database engine is important as well for embedded use. This time is not measured currently. Also, not tested is the time used to create a database and open an existing database. Here, one (wrapper) connection is opened at the start, and for each step a new connection is opened and then closed.
@performance_1275_h2
@performance_1279_h2
#PolePosition Benchmark
@performance_1276_p
@performance_1280_p
#The PolePosition is an open source benchmark. The algorithms are all quite simple. It was developed / sponsored by db4o.
@performance_1277_th
#Test Case
@performance_1281_th
#Test Case #MySQL
@performance_1278_th
@performance_1282_th
#Unit
@performance_1279_th
@performance_1283_th
H2
@performance_1280_th
@performance_1284_th
HSQLDB
@performance_1281_th
@performance_1285_th
MySQL
@performance_1282_td
@performance_1286_td
#Melbourne write
@performance_1283_td
@performance_1287_td
#ms
@performance_1284_td
@performance_1288_td
#369
@performance_1285_td
@performance_1289_td
#249
@performance_1286_td
@performance_1290_td
#2022
@performance_1287_td
@performance_1291_td
#Melbourne read
@performance_1288_td
@performance_1292_td
#ms
@performance_1289_td
@performance_1293_td
#47
@performance_1290_td
@performance_1294_td
#49
@performance_1291_td
@performance_1295_td
#93
@performance_1292_td
@performance_1296_td
#Melbourne read_hot
@performance_1293_td
@performance_1297_td
#ms
@performance_1294_td
@performance_1298_td
#24
@performance_1295_td
@performance_1299_td
#43
@performance_1296_td
@performance_1300_td
#95
@performance_1297_td
@performance_1301_td
#Melbourne delete
@performance_1298_td
@performance_1302_td
#ms
@performance_1299_td
@performance_1303_td
#147
@performance_1300_td
@performance_1304_td
#133
@performance_1301_td
@performance_1305_td
#176
@performance_1302_td
@performance_1306_td
#Sepang write
@performance_1303_td
@performance_1307_td
#ms
@performance_1304_td
@performance_1308_td
#965
@performance_1305_td
@performance_1309_td
#1201
@performance_1306_td
@performance_1310_td
#3213
@performance_1307_td
@performance_1311_td
#Sepang read
@performance_1308_td
@performance_1312_td
#ms
@performance_1309_td
@performance_1313_td
#765
@performance_1310_td
@performance_1314_td
#948
@performance_1311_td
@performance_1315_td
#3455
@performance_1312_td
@performance_1316_td
#Sepang read_hot
@performance_1313_td
@performance_1317_td
#ms
@performance_1314_td
@performance_1318_td
#789
@performance_1315_td
@performance_1319_td
#859
@performance_1316_td
@performance_1320_td
#3563
@performance_1317_td
@performance_1321_td
#Sepang delete
@performance_1318_td
@performance_1322_td
#ms
@performance_1319_td
@performance_1323_td
#1384
@performance_1320_td
@performance_1324_td
#1596
@performance_1321_td
@performance_1325_td
#6214
@performance_1322_td
@performance_1326_td
#Bahrain write
@performance_1323_td
@performance_1327_td
#ms
@performance_1324_td
@performance_1328_td
#1186
@performance_1325_td
@performance_1329_td
#1387
@performance_1326_td
@performance_1330_td
#6904
@performance_1327_td
@performance_1331_td
#Bahrain query_indexed_string
@performance_1328_td
@performance_1332_td
#ms
@performance_1329_td
@performance_1333_td
#336
@performance_1330_td
@performance_1334_td
#170
@performance_1331_td
@performance_1335_td
#693
@performance_1332_td
@performance_1336_td
#Bahrain query_string
@performance_1333_td
@performance_1337_td
#ms
@performance_1334_td
@performance_1338_td
#18064
@performance_1335_td
@performance_1339_td
#39703
@performance_1336_td
@performance_1340_td
#41243
@performance_1337_td
@performance_1341_td
#Bahrain query_indexed_int
@performance_1338_td
@performance_1342_td
#ms
@performance_1339_td
@performance_1343_td
#104
@performance_1340_td
@performance_1344_td
#134
@performance_1341_td
@performance_1345_td
#678
@performance_1342_td
@performance_1346_td
#Bahrain update
@performance_1343_td
@performance_1347_td
#ms
@performance_1344_td
@performance_1348_td
#191
@performance_1345_td
@performance_1349_td
#87
@performance_1346_td
@performance_1350_td
#159
@performance_1347_td
@performance_1351_td
#Bahrain delete
@performance_1348_td
@performance_1352_td
#ms
@performance_1349_td
@performance_1353_td
#1215
@performance_1350_td
@performance_1354_td
#729
@performance_1351_td
@performance_1355_td
#6812
@performance_1352_td
@performance_1356_td
#Imola retrieve
@performance_1353_td
@performance_1357_td
#ms
@performance_1354_td
@performance_1358_td
#198
@performance_1355_td
@performance_1359_td
#194
@performance_1356_td
@performance_1360_td
#4036
@performance_1357_td
@performance_1361_td
#Barcelona write
@performance_1358_td
@performance_1362_td
#ms
@performance_1359_td
@performance_1363_td
#413
@performance_1360_td
@performance_1364_td
#832
@performance_1361_td
@performance_1365_td
#3191
@performance_1362_td
@performance_1366_td
#Barcelona read
@performance_1363_td
@performance_1367_td
#ms
@performance_1364_td
@performance_1368_td
#119
@performance_1365_td
@performance_1369_td
#160
@performance_1366_td
@performance_1370_td
#1177
@performance_1367_td
@performance_1371_td
#Barcelona query
@performance_1368_td
@performance_1372_td
#ms
@performance_1369_td
@performance_1373_td
#20
@performance_1370_td
@performance_1374_td
#5169
@performance_1371_td
@performance_1375_td
#101
@performance_1372_td
@performance_1376_td
#Barcelona delete
@performance_1373_td
@performance_1377_td
#ms
@performance_1374_td
@performance_1378_td
#388
@performance_1375_td
@performance_1379_td
#319
@performance_1376_td
@performance_1380_td
#3287
@performance_1377_td
@performance_1381_td
#Total
@performance_1378_td
@performance_1382_td
#ms
@performance_1379_td
@performance_1383_td
#26724
@performance_1380_td
@performance_1384_td
#53962
@performance_1381_td
@performance_1385_td
#87112
@performance_1382_p
@performance_1386_p
#There are a few problems with the PolePosition test:
@performance_1383_li
@performance_1387_li
#HSQLDB uses in-memory tables by default while H2 uses persistent tables. The HSQLDB version included in PolePosition does not support changing this, so you need to replace poleposition-0.20/lib/hsqldb.jar with a newer version (for example hsqldb-1.8.0.7.jar), and then use the setting hsqldb.connecturl=jdbc:hsqldb:file:data/hsqldb/dbbench2;hsqldb.default_table_type=cached;sql.enforce_size=true in Jdbc.properties.
@performance_1384_li
@performance_1388_li
#HSQLDB keeps the database open between tests, while H2 closes the database (losing all the cache). To change that, use the database URL jdbc:h2:file:data/h2/dbbench;DB_CLOSE_DELAY=-1
@performance_1385_li
@performance_1389_li
#The amount of cache memory is quite important, specially for the PolePosition test. Unfortunately, the PolePosition test does not take this into account.
@performance_1386_h2
@performance_1390_h2
#Application Profiling
@performance_1387_h3
@performance_1391_h3
#Analyze First
@performance_1388_p
@performance_1392_p
#Before trying to optimize performance, it is important to understand where the problem is (what part of the application is slow). Blind optimization or optimization based on guesses should be avoided, because usually it is not an efficient strategy. There are various ways to analyze an application. Sometimes two implementations can be compared using System.currentTimeMillis(). But this does not work for complex applications with many modules, and for memory problems.
@performance_1389_p
@performance_1393_p
#A good tool to measure both memory usage and performance is the <a href="http://www.yourkit.com">YourKit Java Profiler</a> .
@performance_1390_p
@performance_1394_p
#A simple way to profile an application is to use the built-in profiling tool of java. Example:
@performance_1391_p
@performance_1395_p
#Unfortunately, it is only possible to profile the application from start to end. Another solution is to create a number of full thread dumps. To do that, first run <code>jps -l</code> to get the process id, and then run <code>jstack <pid></code> or <code>kill -QUIT <pid></code> (Linux) or press Ctrl+C (Windows).
@performance_1392_h2
@performance_1396_h2
#Database Profiling
@performance_1393_p
@performance_1397_p
#The ConvertTraceFile tool generates SQL statement statistics at the end of the SQL script file. The format used is similar to the profiling data generated when using java -Xrunhprof. As an example, execute the the following script using the H2 Console:
@performance_1394_p
@performance_1398_p
#Now convert the .trace.db file using the ConvertTraceFile tool:
@performance_1395_p
@performance_1399_p
#The generated file <code>test.sql</code> will contain the SQL statements as well as the following profiling data (results vary):
@performance_1396_h2
@performance_1400_h2
#Database Performance Tuning
@performance_1397_h3
@performance_1401_h3
#Virus Scanners
@performance_1398_p
@performance_1402_p
#Some virus scanners scan files every time they are accessed. It is very important for performance that database files are not scanned for viruses. The database engine does never interprets the data stored in the files as programs, that means even if somebody would store a virus in a database file, this would be harmless (when the virus does not run, it cannot spread). Some virus scanners allow to exclude files by suffix. Make sure files ending with .db are not scanned.
@performance_1399_h3
@performance_1403_h3
トレースオプションを使用する
@performance_1400_p
@performance_1404_p
#If the main performance hot spots are in the database engine, in many cases the performance can be optimized by creating additional indexes, or changing the schema. Sometimes the application does not directly generate the SQL statements, for example if an O/R mapping tool is used. To view the SQL statements and JDBC API calls, you can use the trace options. For more information, see <a href="features.html#trace_options">Using the Trace Options</a> .
@performance_1401_h3
@performance_1405_h3
#Index Usage
@performance_1402_p
@performance_1406_p
#This database uses indexes to improve the performance of SELECT, UPDATE and DELETE statements. If a column is used in the WHERE clause of a query, and if an index exists on this column, then the index can be used. Multi-column indexes are used if all or the first columns of the index are used. Both equality lookup and range scans are supported. Indexes are used to order result sets, but only if the condition uses the same index or no index at all. The results are sorted in memory if required. Indexes are created automatically for primary key and unique constraints. Indexes are also created for foreign key constraints, if required. For other columns, indexes need to be created manually using the CREATE INDEX statement.
@performance_1403_h3
@performance_1407_h3
#Optimizer
@performance_1404_p
@performance_1408_p
#This database uses a cost based optimizer. For simple and queries and queries with medium complexity (less than 7 tables in the join), the expected cost (running time) of all possible plans is calculated, and the plan with the lowest cost is used. For more complex queries, the algorithm first tries all possible combinations for the first few tables, and the remaining tables added using a greedy algorithm (this works well for most joins). Afterwards a genetic algorithm is used to test at most 2000 distinct plans. Only left-deep plans are evaluated.
@performance_1405_h3
@performance_1409_h3
#Expression Optimization
@performance_1406_p
@performance_1410_p
#After the statement is parsed, all expressions are simplified automatically if possible. Operations are evaluated only once if all parameters are constant. Functions are also optimized, but only if the function is constant (always returns the same result for the same parameter values). If the WHERE clause is always false, then the table is not accessed at all.
@performance_1407_h3
@performance_1411_h3
#COUNT(*) Optimization
@performance_1408_p
@performance_1412_p
#If the query only counts all rows of a table, then the data is not accessed. However, this is only possible if no WHERE clause is used, that means it only works for queries of the form SELECT COUNT(*) FROM table.
#When executing a query, at most one index per joined table can be used. If the same table is joined multiple times, for each join only one index is used. Example: for the query SELECT * FROM TEST T1, TEST T2 WHERE T1.NAME='A' AND T2.ID=T1.ID, two index can be used, in this case the index on NAME for T1 and the index on ID for T2.
@performance_1411_p
@performance_1415_p
#If a table has multiple indexes, sometimes more than one index could be used. Example: if there is a table TEST(ID, NAME, FIRSTNAME) and an index on each column, then two indexes could be used for the query SELECT * FROM TEST WHERE NAME='A' AND FIRSTNAME='B', the index on NAME or the index on FIRSTNAME. It is not possible to use both indexes at the same time. Which index is used depends on the selectivity of the column. The selectivity describes the 'uniqueness' of values in a column. A selectivity of 100 means each value appears only once, and a selectivity of 1 means the same value appears in many or most rows. For the query above, the index on NAME should be used if the table contains more distinct names than first names.
@performance_1412_p
@performance_1416_p
#The SQL statement ANALYZE can be used to automatically estimate the selectivity of the columns in the tables. This command should be run from time to time to improve the query plans generated by the optimizer.
@performance_1413_h3
@performance_1417_h3
#Optimization Examples
@performance_1414_p
@performance_1418_p
#See <code>src/test/org/h2/samples/optimizations.sql</code> for a few examples of queries that benefit from special optimizations built into the database.
@performance_1415_h3
@performance_1419_h3
#Cache Size and Type
@performance_1416_p
@performance_1420_p
#By default the cache size of H2 is quite small. Consider using a larger cache size, or enable the second level soft reference cache. See also <a href="features.html#cache_settings">Cache Settings</a> .
