changelog.html

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<!--
Copyright 2004-2011 H2 Group. Multiple-Licensed under the H2 License, Version 1.0,
and under the Eclipse Public License, Version 1.0
(http://h2database.com/html/license.html).
Initial Developer: H2 Group
-->
<html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en">
<head><meta http-equiv="Content-Type" content="text/html;charset=utf-8" /><title>
Change Log
</title><link rel="stylesheet" type="text/css" href="stylesheet.css" />
<!-- [search] { -->
<script type="text/javascript" src="navigation.js"></script>
</head><body onload="frameMe();">
<table class="content"><tr class="content"><td class="content"><div class="contentDiv">
<!-- } -->

<h1>Change Log</h1>

<h2>Next Version (unreleased)</h2>
<ul><li>-
</li></ul>

<h2>Version 1.3.165 (2012-03-18)</h2>
<ul><li>Better string representation for decimal values (for example 0.00000000 instead of 0E-26).
</li><li>Prepared statements could only be re-used if the same data types were used
    the second time they were executed.
</li><li>In error messages about referential constraint violation, the values are now included.
</li><li>SCRIPT and RUNSCRIPT: the password can now be set using a prepared statement.
    Previously, it was required to be a literal in the SQL statement.
</li><li>MySQL compatibility: SUBSTR with a negative start index now works like MySQL.
</li><li>When enabling autocommit, the transaction is now committed (as required by the JDBC API).
</li><li>The shell script <code>h2.sh</code> did not work with spaces in the path.
    It also works now with quoted spaces in the argument list.
    Thanks a lot to Shimizu Fumiyuki for the patch!
</li><li>If the transaction log could not be truncated because of an uncommitted transaction,
    now "Transaction log could not be truncated" is written to the .trace.db file.
    Before, the database file was growing and it was hard to find out what the root cause was.
    To avoid the database file from growing, a new feature to automatically rollback the oldest
    transaction is available now. To enable it, append ;LOG_SIZE_LIMIT=32 to the database URL
    (in that case, the oldest session is rolled back if the transaction log is 32 MB).
</li><li>ALTER TABLE ADD can now add more than one column at a time.
</li><li>Issue 380: ALTER TABLE ADD FOREIGN KEY with an explicit index didn't verify
    the index can be used, which would lead to a NullPointerException later on.
</li><li>Issue 384: the wrong kind of exception (NullPointerException) was thrown in a UNION query
    with an incorrect ORDER BY expression.
</li><li>Issue 362: support LIMIT in UPDATE statements.
</li><li>Browser: if no default browser is set, Google Chrome is now used if available.
    If not available, then Konqueror, Netscape, or Opera is used if available (as before).
</li><li>CSV tool: new feature to disable writing the column header
    (option writeColumnHeader).
</li><li>CSV tool: new feature to preserve the case sensitivity of column names
    (option caseSensitiveColumnNames).
</li><li>PostgreSQL compatibility: LOG(x) is base 10 in the PostgreSQL mode.
</li></ul>

<h2>Version 1.3.164 (2012-02-03)</h2>
<ul><li>New built-in function ARRAY_CONTAINS.
</li><li>Some DatabaseMetaData methods didn't work when using ALLOW_LITERALS NONE.
</li><li>Trying to convert a VARCHAR to UUID will now fail if the text contains
    a character that is not a hex digit, '-', or not a whitespace.
</li><li>TriggerAdapter: in "before" triggers, values can be changed using the ResultSet.updateX methods.
</li><li>Creating a table with column data type NULL now works (even if not very useful).
</li><li>ALTER TABLE ALTER COLUMN no longer copies the data for widening conversions
    (for example if only the precision was increased) unless necessary.
</li><li>Multi-threaded kernel: concurrently running an online backup
    and updating the database resulted in a broken (transactionally incorrect)
    backup file in some cases.
</li><li>The script created by SCRIPT DROP did not always work if multiple views
    existed that depend on each other.
</li><li>MathUtils.getSecureRandom could log a warning to System.err in case
    the /dev/random is very slow, and the System.getProperties().toString()
    returned a string larger than 64 KB.
</li><li>The database file locking mechanism "FS" (;FILE_LOCK=FS)
    did not work on Linux since version 1.3.161.
</li><li>Sequences: the functions NEXTVAL and CURRVAL did not work as expected
    when using quoted, mixed case sequence names.
</li><li>The constructor for Csv objects is now public, and Csv.getInstance() is now deprecated.
</li><li>SimpleResultSet: updating a result set is now supported.
</li><li>Database URL: extra semicolons are not supported.
</li></ul>

<h2>Version 1.3.163 (2011-12-30)</h2>
<ul><li>On out of disk space, the database could get corrupt sometimes,
    if later write operations succeeded. The same problem happened on other
    kinds of I/O exceptions (where one or some of the writes fail, but subsequent writes
    succeed). Now the file is closed on the first unsuccessful
    write operation, so that later requests fail consistently.
</li><li>DatabaseEventListener.diskSpaceIsLow() is no longer supported because
    it can't be guaranteed that it always works correctly.
</li><li>XMLTEXT now supports an optional parameter to escape newlines.
</li><li>XMLNODE now support an optional parameter to disable indentation.
</li><li>Csv.write now formats date, time, and timestamp values using java.sql.Date / Time / Timestamp.toString().
    Previously, ResultSet.getString() was used, which didn't work well for Oracle.
</li><li>The shell script <code>h2.sh</code> can now be run from within a different directory.
    Thanks a lot to Daniel Serodio for the patch!
</li><li>The page size of a persistent database can now be queries using:
    select * from information_schema.settings where name = 'info.PAGE_SIZE'
</li><li>In the server mode, BLOB and CLOB objects are no longer closed when the result set is closed
    (as required by the JDBC spec).
</li></ul>

<h2>Version 1.3.162 (2011-11-26)</h2>
<ul><li>The following system properties are no longer supported:
    <code>h2.allowBigDecimalExtensions</code>,
    <code>h2.emptyPassword</code>,
    <code>h2.minColumnNameMap</code>,
    <code>h2.returnLobObjects</code>,
    <code>h2.webMaxValueLength</code>.
</li><li>When using a VPN, starting a H2 server did not work (for some VPN software).
</li><li>Oracle compatibility: support for DECODE(...).
</li><li>Lucene fulltext search: creating an index is now faster if the table already contains data.
    Thanks a lot to Angel Leon from the FrostWire Team for the patch!
</li><li>Update statements with a column list in brackets did not work if the list only contains one column.
    Example: update test set (id)=(id).
</li><li>Read-only databases in a zip file did not work when using the -baseDir option.
</li><li>Issue 334: SimpleResultSet.getString now also works for Clob columns.
</li><li>Subqueries with an aggregate did not always work. Example:
    select (select count(*) from test where a = t.a and b = 0) from test t group by a
</li><li>Server: in some (theoretical) cases, exceptions while closing the connection were ignored.
</li><li>Server.createTcpServer, createPgServer, createWebServer: invalid arguments are now detected.
</li><li>The selectivity of LOB columns is no longer calculated
    because indexes on LOB columns are not supported
    (however this should have little effect on performance, as the selectivity
    is calculated from the hash code and not the data).
</li><li>New experimental system property "h2.modifyOnWrite":
    when enabled, the database file is only modified when writing to the database.
    When enabled, the serialized file lock is much faster for read-only operations.
</li><li>A NullPointerException could occur in TableView.isDeterministic for invalid views.
</li><li>Issue 180: when deserializing objects, the context class loader is used
    instead of the default class loader if the system property "h2.useThreadContextClassLoader" is set.
    Thanks a lot to Noah Fontes for the patch!
</li><li>When using the exclusive mode, LOB operations could cause the thread to block.
    This also affected the CreateCluster tool (when using BLOB or CLOB data).
</li><li>The optimization for "group by" was not working correctly if the group by column
    was aliased in the select list.
</li><li>Issue 326: improved support for case sensitive (mixed case) identifiers
    without quotes when using DATABASE_TO_UPPER=FALSE.
</li></ul>

<h2>Version 1.3.161 (2011-10-28)</h2>
<ul><li>Issue 351: MySQL mode: can not create a table with the column "KEY",
    and can not open databases where such a table already exists.
</li><li>TCP server: when using the trace option ("-trace"), the trace output contained
    unnecessary stack traces when stopping the server.
</li><li>Issue 354: when using the multi-threaded kernel option,
    and multiple threads concurrently prepared SQL statements that use the same view
    or the same index of a view, then in some cases an infinite loop could occur.
</li><li>Issue 350: when using instead of triggers, executeUpdate for delete operations always returned 0.
</li><li>Some timestamps with timezone were not converted correctly.
    For example, in the PST timezone, the timestamp 2011-10-26 08:00:00Z was converted to
    2011-10-25 25:00:00 instead of 2011-10-26 01:00:00.
    Depending on the database operation, this caused subsequent error.
</li><li>Sequences with cache size smaller than 0 did not work correctly.
</li><li>Issue 356: for Blob objects, InputStream.skip() returned 0,
    causing EOFException in Blob.getBytes(.., ..).
</li><li>Updatable result sets: if the value is not set when inserting a new row,
    the default value is now used (the same behavior as MySQL, PostgreSQL, and Apache Derby)
    instead of NULL.
</li><li>Conditions of the form IN(SELECT ...) where slow and increased the database size
    if the subquery result size was larger then the configured MAX_MEMORY_ROWS.
</li><li>Issue 347: the RunScript and Shell tools now ignore empty statements.
</li><li>Issue 246: improved error message for data conversion problems.
</li><li>Issue 344: the build now supports a custom Maven repository location.
</li><li>Issue 334: Java functions that return CLOB or BLOB objects could not be used as tables.
</li><li>SimpleResultSet now supports getColumnTypeName and getColumnClassName.
</li><li>SimpleResultSet now has minimal BLOB and CLOB support.
</li><li>Improved performance for large databases (many GB), and databases with a small page size.
</li><li>The Polish translation has been improved by Jaros&#322;aw Kokoci&#324;ski.
</li><li>When using the built-in connection pool, after calling the "shutdown" SQL statement,
    a warning was written to the .trace.db file about an unclosed connection.
</li><li>Improved compatibility for "fetch first / next row(s)". Thanks a lot to litailang for the patch!
</li><li>Improved compatibility with the Java 7 FileSystem abstraction.
</li></ul>

<h2>Version 1.3.160 (2011-09-11)</h2>
<ul><li>Computed columns could not refer to itself.
</li><li>Issue 340: Comparison with "x = all(select ...)" or similar in a view or
    subquery that was used as a table returned the wrong result.
</li><li>Comparison with "x = all(select ...)" or similar returned the wrong result for some cases
    (for example if the subquery returned no rows, or multiple rows with the same value, or null, or if x was null).
</li><li>Issue 335: Could not run DROP ALL OBJECTS DELETE FILES on older databases with CLOB or BLOB data.
    The problem was that the metadata table was not locked in some cases, so that
    a rollback could result in a corrupt database in a database if the lob storage was upgraded.
</li><li>Source code switching using //## has been simplified.
    The Java 1.5 tag has been removed because is no longer needed.
</li><li>The JDBC methods PreparedStatement.setTimestamp, setTime, and setDate with a calendar,
    and the methods ResultSet.getTimestamp, getTime, and getDate with a calendar converted
    the value in the wrong way, so that for some timestamps the converted value was wrong
    (where summertime starts, one hour per year).
</li><li>Invalid tables names in 'order by' columns were not detected in some cases. Example:
    select x from dual order by y.x
</li><li>Issue 329: CASE expression: data type problem is not detected.
</li><li>Issue 311: File lock mode serialized: selecting the next value from a sequence didn't work
    after a pause, because the database thought this is a read-only operation.
</li></ul>

<h2>Version 1.3.159 (2011-08-13)</h2>
<ul><li>Creating a temporary table with the option 'transactional' will now also create the indexes
    in transactional mode, if the indexes are included in the 'create table' statement as follows:
    "create local temporary table temp(id int primary key, name varchar, constraint x index(name)) transactional".
</li><li>The database file size grows now 35%, but at most 256 MB at a time.
</li><li>Improved error message on network configuration problems.
</li><li>The build now support an offline build using ./build.sh offline.
    This will list the required dependencies if jar files are missing.
</li><li>The BLOB / CLOB data was dropped a little bit before the table was dropped.
    This could cause "lob not found" errors when the process was killed while a table was dropped.
</li><li>"group_concat(distinct ...)" did not work correctly in a view or subquery (the 'distinct' was lost). Example:
    select * from (select group_concat(distinct 1) from system_range(1, 3));
</li><li>Database URLs can now be re-mapped to another URL using the system property
    "h2.urlMap", which points to a properties file with database URL mappings.
</li><li>When using InputStream.skip, trying to read past the end of a BLOB failed with
    the exception "IO Exception: Missing lob entry: ..." [90028-...].
</li><li>The in-memory file system "memFS:" now has limited support for directories.
</li><li>To test recovery, append ;RECOVER_TEST=64 to the database URL.
    This will simulate an application crash after each 64 writes to the database file.
    A log file named databaseName.h2.db.log is created that lists the operations.
    The recovery is tested using an in-memory file system, that means it may require a larger heap setting.
</li><li>Converting a hex string to a byte array is now faster.
</li><li>The SQL statement "shutdown defrag" could corrupt the database if the process was killed
    while the shutdown was in progress. The same problem could occur when the database
    setting "defrag_always" was used.
</li></ul>

<h2>Version 1.3.158 (2011-07-17)</h2>
<ul><li>Fulltext search (native): after re-opening the connection once the fulltext index was created,
    a transaction rollback did not roll back the modifications in the index.
</li><li>Support for ROW_NUMBER() OVER() as an alias for ROWNUM()
    for compatibility with PostgreSQL and Apache Derby.
</li><li>Issue 325: a deadlock in the fulltext search could occur if the connection that was
    used while initializing the fulltext search was later used to query the database.
    This was actually not only a problem for fulltext search, but also for other triggers
    that used the initializing connection later on.
</li><li>DATABASE_TO_UPPER: when set to false, all identifier names (table names, column names)
    are case sensitive (except aggregate, built-in functions, data types, and keywords).
    This is for improved compatibility with MySQL and PostgreSQL.
</li><li>When upgrading from an older 1.3.x version to version 1.3.157, when using BLOB or CLOB data,
    the database could only be opened once.
</li><li>SUM on a boolean expression will now count the number of 'true' rows.
</li><li>Simplified test cases using assertThrows(...).
</li><li>A Java function that returns a result set that was generated using the provided connection
    could cause an internal error in some cases. Now such result sets are copied early on.
</li><li>Data type RESULT_SET: information_schema.function_aliases returned type 0; now type -10 is returned (the same as Oracle: OracleTypes.CURSOR = -10).
</li><li>Data type RESULT_SET: tables with columns of type RESULT_SET could be created, but no data could be persisted in this column.
</li><li>H2 Console: improved system tray icon for Mac OS X (transparent background).
</li><li>String.toUpperCase() was used a few places, which is problematic when using the Turkish locale.
    The method has been replaced with toUpperCase(Locale.ENGLISH) to solve such problems.
</li><li>Shell tool: the built-in command "distinct" has been removed
    (use the SQL statements "show tables" / "show columns from tableName" instead).
    The result set formatting has been improved.
</li><li>MERGE: if a unique key was violated (but not the primary key or the key columns of the merge itself),
    the wrong exception was thrown.
</li><li>IBM DB2 and Apache Derby compatibility: support the pseudo-table SYSIBM.SYSDUMMY1.
</li><li>ROWNUM() did not work in combination with IN(..). The following query did not work as expected:
    select * from (select rownum r from test) where r in (1, 2).
</li><li>H2 Console autocomplete: the autocomplete feature didn't support quoted names.
</li><li>It is now longer allowed to create an index on a CLOB or BLOB column
    (except for in-memory databases or indexes), because recovery doesn't work
    on such columns. Fulltext indexes on such column are still supported of course.
</li><li>The PUBLIC schema could be renamed, which could result in a database that
    can't be opened normally. Now it can no longer be renamed.
    The schema INFORMATION_SCHEMA can also no longer be renamed.
</li><li>Server mode: the problem with server and client using different
    timezones with different daylight saving rules has been solved
    (this only applies when both the client and the server use this or a newer version).
</li><li>There is a new system property "h2.storeLocalTime" to avoid
    problems opening a database file in a timezone that has different
    daylight saving rules (see also below).
</li><li>Date, time, and timestamp data type processing has been re-implemented.
    Time now supports nanoseconds resolution and now now supports a wider range
    (negative and large values), similar to PostgreSQL.
</li><li>SQL statements with a non-breaking space were considered invalid.
</li></ul>

<h2>Version 1.3.157 (2011-06-25)</h2>
<ul><li>The syntax for CREATE TABLE ... AS SELECT ... has been changed:
    options such as NOT PERSISTENT are now before the query (AS SELECT), because otherwise
    they might be interpreted as a part of the query.
</li><li>With version 1.3.156, the DB2 mode could not be used with CLOB and BLOB data.
</li><li>When upgrading from an older 1.3.x version to version 1.3.156,
    CLOB and BLOB data was lost in many cases.
</li><li>The optimization for COUNT(..) on columns that are not nullable
    was also used for COUNT(DISTINCT ..), which is incorrect.
</li><li>PgServer: non-admin users could not open a database.
</li><li>Non-admin users could not open a database using a mode (MODE=xxx)
    if the database was already open, even if the mode matched.
</li><li>The SQL state of the following exceptions has been changed:
    From 90009, 90010, 90011 to 22007: "Cannot parse DATE/TIME/TIMESTAMP constant ...".
</li><li>There is a problem when opening a database file in a timezone that has different
    daylight saving rules: the time part of dates where the daylight saving doesn't match
    will differ. This is not a problem within regions that use the same rules (such as, within
    USA, or within Europe), even if the timezone itself is different. As a workaround, export the
    database to a SQL script using the old timezone, and create a new database in the new
    timezone. The same problem occurs when accessing a database over TCP/IP if the client and server
    timezones use different rules, if an old server or client is used.
</li></ul>

<h2>Version 1.3.156 (2011-06-17)</h2>
<ul><li>Issue 323: NullPointerException when using IN(...) with a function table.
</li><li>Clob.toString() and Blob.toString() now include the trace identifier.
</li><li>The SQL parser silently ignored characters such as '^' or '\'.
    Now a syntax error is thrown.
</li><li>ROUND(..) now also works with just one parameter.
</li><li>Access to system tables is now more restrictive for non-admin users
    (the tables can be still listed, but some of the data is not included).
</li><li>BLOB: InputStream.skip is now more efficient in embedded mode.
</li><li>The default expression of a column may no longer reference the table,
    because a database with such a table couldn't be opened normally.
</li><li>Referenced objects in check constraints could be dropped, which resulted
    in a database that can't be opened normally.
</li><li>Improved Oracle compatibility for NVL2. Thanks again to litailang for the patch.
</li><li>CSVREAD now supports the option 'preserveWhitespace'.
</li><li>Recursive queries with many rows could throw an IndexOutOfBoundsException.
</li><li>The auto-server mode can't be combined with an in-memory database.
    This invalid combination wasn't detected so far.
    Now trying to open a database in this way fails.
</li><li>Improved performance for reading and writing date and time values.
</li><li>Java functions: array component types are now preserved, so that a ResultSet.getObject()
    will return Integer[] if the Java functions returns Integer[]. Thanks to Noel Grandin for the patch.
    Component types are transferred over the network, but not persisted in the database file as of now.
</li><li>Recover tool: the script generated by the recover tool didn't work
    if fulltext search was used, because the triggers were created before the primary keys.
</li><li>Issue 321: the database does not expect XA rollback without XA prepare
    (an exception was unnecessarily written into the .trace.db file).
</li><li>Read-only databases: some queries (such as distinct queries)
    tried to create temporary tables, which failed for read only databases.
</li><li>This database no longer uses finalize() except for temporary files.
    The system property "h2.runFinalize" is no longer supported.
    Unclosed connections are still detected, but the opening stack trace is now collected
    only if connections are not closed repeatedly (starting with the second unclosed connection).
    To detect unclosed connections, a PhantomReference is used.
</li><li>Issue 322: Left joins didn't work when used as subquery in the form clause.
</li><li>A prepared statement if type CALL that returned a result set could
    throw a NullPointerException if executed multiple times.
</li><li>The following built-in functions were marked as deterministic and therefore
    not re-executed when using in a prepared statement:
    LOCK_TIMEOUT, LINK_SCHEMA, CANCEL_SESSION, FILE_READ.
</li><li>Count on a column that can not be null is now optimized to COUNT(*).
    Thanks a lot to Ross Judson for providing a patch and test case!
</li><li>The following prepared statements are now cached as well:
    insert, update, delete, call, merge, and transactional commands.
</li><li>The built-in connection pool is now a bit faster.
</li><li>The scale was not set correctly in some cases when using
    CREATE TABLE AS SELECT if there was no explicit column definition.
</li><li>When trying to connect to a server using TCP/IP failed, it will retry at most
    until the timeout. Before, it was the timeout multiplied with the retry count.
</li></ul>

<h2>Version 1.3.155 (2011-05-27)</h2>
<ul><li>When altering a table with triggers, new triggers are created
    and initialized with a table name that ends with tableName_COPY_x_y.
    The plan is that this will not be required in future versions.
</li><li>If some of the file system operations failed, System.gc() was called in a loop.
    Now it is called once at most.
</li><li>Triggers (including the fulltext search) could throw the exception
    'Unexpected code path' in Session.log.
</li><li>CURRENT_TIMESTAMP() and so on now return the same value within a transaction.
</li><li>The statement CALL no longer converts an ARRAY return value to a list of values.
    Now an ARRAY is returned.
</li><li>VALUES is now supported as a standalone command and as a table source:
    SELECT * FROM (VALUES(1, 'Hello'), (2, 'World')) AS V;
</li><li>TIMESTAMPADD is now an alias for DATEADD.
    Most SQL_TSI_ constants are now supported for for TIMESTAMPADD and TIMESTAMPDIFF
    (all except SQL_TSI_QUARTER and SQL_TSI_FRAC_SECOND).
</li><li>Issue 313: NullPointerException in select query with a subquery or view.
</li><li>Native fulltext search: the characters '&lt;', '&gt;', and '\' are now also whitespace
    characters. Also, the list of whitespace characters can be changed using
    FullText.setWhitespaceChars(conn, ...)
</li><li>When reading from the classpath (for example read_file('classpath:logo.png')),
    now the content class loader is used if the resource can't be found otherwise.
</li><li>The statement "script drop" create a script that couldn't be executed if
    a table contained a reference to a user defined function.
</li><li>Opening a database with both multi-threaded mode and lock mode 0 (disable locking)
    was possible (but only once), even thought it does not work properly.
    Now it is no longer allowed.
</li><li>LOB objects are now deleted (and the empty space is reused) when the transaction is committed.
    So far, the space wasn't re-used until the connection was closed.
</li><li>SCRIPT DROP did not use the cascade option for table, which would cause problems
    if a view was dropped before a table it depends on.
    Now the cascade option is used.
</li><li>The query plan of queries with subqueries that contain nested joins was wrong.
    This could result in syntax exceptions for correct queries. Example:
    select 1 from (select 2 from ((test t1 inner join test t2 on t1.id=t2.id) inner join test t3 on t3.id=t1.id)) x;
</li><li>DatabaseEventListener.init was called with a null url when using the server mode.
</li><li>Support for the function LN (an alias for LOG).
</li><li>Support for the function CEIL (an alias for CEILING).
</li><li>Issue 315: Access to LOBs could cause a Java level deadlock.
</li><li>Support for the ICU4J collator.
</li><li>Improved Oracle compatibility: support for NVL2. Thanks to litailang for the patch!
</li><li>Improved PostgreSQL compatibility: support for RANDOM() in addition to RAND().
</li><li>There was a classloader memory leak problem because a class contained a static
    references to an exception (including stack trace).
</li><li>Split file system: truncating a file now deletes the parts in reverse order,
    so that the file list is consistent if the process is interrupted while truncating.
</li><li>The H2 JDBC client can now be used in an unsigned Applet.
    The problem was that System.getProperty throws a SecurityException, which is now ignored.
</li><li>The condition "in(select ...)" did not work correctly if the subquery
    could not be converted to a "distinct" query, as in:
    "select * from dual where x in (select x from dual group by x order by max(x))".
    Now the subquery needs to be made distinct manually if this optimization is needed.
</li><li>In the last release, the error code for "Wrong user name or password" was still
    08004 instead of 28000, which resulted in the wrong error message.
</li><li>In some cases, creating a new table or altering an existing table threw the exception:
    Unique index or primary key violation: "PRIMARY KEY ON """".PAGE_INDEX".
</li><li>The Shell tool no longer supports the built-in command "show",
    because it is a legal SQL statement for H2 and MySQL.
    However, the "describe" command is still supported, and now lists all tables
    if called without parameter.
</li><li>The sorted insert mode (as in "insert into ... direct sorted select")
    did not work for not-persisted tables in a persisted database.
    It threw a ClassCastException.
</li><li>Issue 308: Statement.getGeneratedKeys() now returns an empty result set if no key was generated.
</li><li>The h2small.jar (created with build jarSmall) included the Android API.
    This has been removed, shrinking the jar file by 21 KB.
</li><li>When creating a table, the precision must now be at least as large as the scale.
</li><li>Support for Java 1.3 and Java 1.4 has been removed.
</li><li>Improved error message for syntax error: the list of expected tokens was sometimes not set correctly.
</li><li>Database file growth can now be limited using the database setting PAGE_STORE_MAX_GROWTH.
    Slower growth may slow down operation, but speed up closing database.
</li><li>A read only database with writable linked tables could lead to a DATABASE_IS_READ_ONLY (90097) error.
    This is fixed now.
</li></ul>

<h2>Version 1.3.154 (2011-04-04)</h2>
<ul><li>The SQL state of the following exceptions has been changed:
    From 90005 to 22001: "Value too long for column ...: ...".
    From 90039 to 22003: "The value is too large for the precision ..." is merged with "Numeric value out of range".
    From 90110 to 22003: "... out of range" is merged with "Numeric value out of range".
    From 90021 to 22018: "Data conversion error converting ...".
    From 90006 to 23502: "NULL not allowed for column ...".
    From 23003 to 23503: "Referential integrity constraint violation: ...".
    From 23001 to 23505: "Unique index or primary key violation: ...".
    From 23002 to 23506: "Referential integrity constraint violation: ...".
    From 90056 to 23507: "No default value is set for column ...".
    From 23000 to 23513: "Check constraint violation: ...".
    From 08004 to 28000: "Wrong user name or password".
    From 90051 to 57014: "Statement was canceled or the session timed out".
</li><li>The H2 Console now detects unsupported command line options.
    (In the future, it will be combined with the Server tool.)
</li><li>MySQL compatibility: SHOW COLUMN FOR schemaName.columnName.
</li><li>DatabaseMetaData.getCrossReference now also lists the name of the primary index (PK_NAME).
</li><li>If the database contained many CLOB or BLOB values, then closing a connection was very slow.
    Reason: was that when closing a connection, the temporary CLOB and BLOB values for this connection
    were removed, but there was no index on this column.
    For existing databases with many CLOB and BLOB values, creating the index may take a few seconds.
    The index is added to existing databases automatically.
</li><li>If a CLOB or BLOB was copied into the same table using INSERT INTO X ... SELECT ... FROM X,
    and then the original row was deleted, then the copied value was also deleted.
    This could also result in an ArrayIndexOutOfBoundsException on checkpoint or when closing the database.
</li><li>Improved performance (up to 10%) due to changed synchronization in the page store.
</li><li>Reading a resource from the classpath didn't work if there was a leading slash
    (classpath:/org/...). Now resources can be loaded with or without leading slash.
</li><li>Support INSERT INTO TEST SET ID = 1, NAME = 'World' (MySQL compatibility).
</li><li>Issue 304: The condition [NOT] IN (SELECT ...) could throw the exception "Unexpected code path" if the subquery contained ORDER BY.
</li><li>ALTER TABLE ALTER ADD / REMOVE /ALTER COLUMN dropped some dependent objects
    (access rights, triggers) of views that depend on the modified table.
</li><li>CREATE OR REPLACE VIEW dropped some dependent objects (access rights, triggers)
    if the view already existed before.
</li><li>New aggregate function HISTOGRAM to calculate the histogram.
    The plan is to use this function is only used internally when running ANALYZE
    to generate the histogram that is then used by the query optimizer.
</li><li>New system function TRUNCATE_VALUE to truncate a value to the required precision.
    This is the same as for CAST, but without changing the data type, and support for truncating decimal values.
    (The main use case for this function is to build a histogram of the truncated values).
    Now BINARY, BLOB, and CLOB can be truncated / cast efficiently as well.
</li><li>The precision of an ARRAY value is now the sum of the precision of all elements,
    and result set values have the precision Integer.MAX_VALUE.
</li><li>CAST(x AS VARCHAR_IGNORECASE(y) / CHAR(y)) returned a VARCHAR
    if the new precision was lower than the old precision.
</li><li>SHUTDOWN DEFRAG didn't defragment the database completely if there was a unique hash index for a memory table.
</li><li>There were some smaller changes in the encrypting file system (work in progress).
</li><li>Some time and timezone functions didn't support years before 1.
</li></ul>

<h2>Version 1.3.153 (2011-03-14)</h2>
<ul><li>Issue 294: The Maven source bundle now contains a manifest file that allows Eclipse to automatically
    attach the source code (if Maven is used).
</li><li>A query with an explicit LIMIT 0 will now return no rows (so far it meant no limit),
    which is compatible with PostgreSQL and MySQL.
    A negative limit value, (as well as LIMIT NULL) mean no limit (so far a negative limit meant a limit of one row).
    There were similar problems with OFFSET.
</li><li>A new encrypting file system has been implemented.
    It is currently experimental, and not included in the main jar file.
    The idea is to allow encrypting all files, not just database files.
    Also implemented is a simple file system shell tool. At some point,
    this might replace the Backup, ChangeFileEncryption, DeleteDbFiles, and Restore tools.
</li><li>Lob in database: storing lob objects was not correctly synchronized.
    This was specially a problem when using Connection.createBlob() / createClob().
</li><li>Updated the license page (thanks a lot to Abe to help me with this).
</li><li>Support for database paths with '\' on non-Windows systems.
</li><li>The NIO cleaner hack wasn't working as expected, because the clear method was called
    instead of the clean method. This has been fixed. Please note this hack is not enabled by default,
    to try it you need to set the system property "h2.nioCleanerHack" to true.
</li><li>Improved MS SQL Server and MySQL compatibility: support DELETE with TOP or LIMIT.
</li><li>Support for COSH, SINH, and TANH functions.
</li><li>Support for the % operator (modulo) thanks to Noel Grandin.
</li><li>Issue 288: Some right outer join queries failed to produce the correct result or
    threw exceptions such as "column x must be in the group by list".
</li><li>Some right outer joins with invalid column referenced (typos) threw a NullPointerException instead of
    "column not found" exception.
</li><li>New database setting EARLY_FILTER to allow table implementations
    to apply filter conditions early on.
</li><li>Cluster: in a two node cluster, if cluster node stopped, and autocommit was enabled,
    then changes on the remaining cluster node didn't get committed automatically.
</li><li>Issue 291: Allow org.h2.value.Value as FUNCTION ALIAS params and return value.
</li><li>Issue 265: Linked tables: auto-reconnect if the backside connection is lost
    (workaround for the MySQL problem that disconnects after 8 hours of inactivity).
</li><li>Linked tables: the index conditions was sometimes not used when querying the remote database.
</li><li>Issue 294: OSGi: the versions were missing in manifest package exports.
</li><li>The option -baseDir didn't work with symbolic links.
</li><li>Support for ALTER TABLE ADD COLUMN IF NOT EXISTS.
</li><li>Database-level connection settings could only be set in the database URL,
    but not using the Properties parameter of DriverManager.getConnection(String url, Properties info).
</li></ul>

<h2>Version 1.3.152 Beta (2011-03-01)</h2>
<ul><li>H2 Console: browser detection didn't work for Mac OS X with OpenJDK 1.7.
</li><li>The double or float value -0.0 is now distinct from 0.0 (as in Java).
</li><li>TCP and PG server: if the default port is already in use, the services are now started on
    another free port (unless the port is explicitly set). This allows to start the H2 Console
    even if the TCP or PG servers are already running.
</li><li>LIKE didn't use an index unless the right hand side was a literal or parameter.
    Now an index is used for LIKE CONCAT(?, '%').
</li><li>Array literals with zero or one element were not be parsed correctly.
    Now trailing commas are ignored (as in Java). Expressions of the form '(1)' are still parsed as the value 1.
    An empty array is '()'. An array with one element must contain a comma to be parsed as an array, as in '(1,)'.
</li><li>Issue 290: Conditions using subqueries were sometimes evaluated before much simpler conditions,
    which was very bad for performance. This includes subqueries returning a value, EXISTS subqueries,
    and IN(SELECT ..) subqueries.
</li><li>Issue 288: Some queries with right outer joins or nested joins could throw a NullPointerException.
</li><li>The database can now be compiled with Java 7.
    Stub methods for the new JDBC 4.1 methods have been implemented,
    but the new features can't be used yet.
</li><li>Issue 289: Railroads (BNF) for "Column Definition" was incorrect.
</li><li>Issue 287: Railroads (BNF) is not readable for certain web-browser configurations (dark background).
</li><li>Executing a CSVREAD query multiple times could result in an exception if the data was changed on disk.
</li><li>SQL railroad diagrams: improved images for people with black background settings.
</li><li>Build-in documentation search: improved results if searching for multiple words.
</li><li>DatabaseMetaData.getColumns: there is a typo in the JDBC specification, so that the column #19 is named
    SCOPE_CATLOG instead of SCOPE_CATALOG. For compatibility with the specification, a new column #24
    named SCOPE_CATLOG is added. This typo is fixed in the JDBC 4.1 specification.
    Please note that MySQL only supports SCOPE_CATALOG, while other databases only support SCOPE_CATLOG.
</li><li>Build: jar files are now downloaded using Maven if possible (and cached in the local Maven repository).
    Unfortunately, in many cases this will download more data from the Maven repository.
</li><li>JaQu has been improved a lot thanks to the new committer James Moger.
</li><li>New sample application that shows how to use "instead of" triggers to support updatable views.
</li><li>H2 Console: opening a MS SQL Server database is now faster.
    Also, listing the database meta data sometimes resulted in an exception.
</li><li>Linked tables / MS SQL Server: if the tables were created in mixed case in MS SQL Server,
    then the identifiers in the linked tables were case sensitive (in H2).
</li><li>MVCC: the probability of a lock timeout is now lower if multiple threads try to update the same rows.
</li><li>Building only the documentation (without compiling all classes) didn't work, specially: ./build.sh clean javadocImpl.
</li><li>Documentation on how data is stored internally and how indexes work (in the performance section).
</li><li>Some people reported NullPointerException in FileObjectDiskMapped.
    The most likely explanation is that multiple threads access the same object at the same time.
    Therefore, the public methods in this class are now synchronized.
</li><li>Improved error detection when starting a server with invalid arguments,
    such as "-tcpPort=9091" or "-tcpPort 9091" (as one parameter) instead of "-tcpPort", "9091".
</li><li>The function STRINGDECODE ignored characters after a non-escaped double quote.
    This is no longer the case.
</li><li>The new CSV option feature introduced in version 1.3.150 didn't support spaces.
    Now the escape mechanism is changed to support it.
</li><li>The pseudo-column "_ROWID_" now supports insert, update, and merge (only enabled for version 1.3.x).
</li><li>Deleting rows in the trigger table inside a trigger could result in a NullPointerException.
</li></ul>

<!-- [close] { --></div></td></tr></table><!-- } --><!-- analytics --></body></html>