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!

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

The selectivity of LOB columns is no longer calculated because indexes on LOB columns are not supported.

The selectivity of LOB columns is no longer calculated because indexes on LOB columns are not supported.

New experimental system property "h2.modifyOnWrite": when enabled, the database file is only modified when writing to the database. This should speed up the serialized file lock.

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!

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!

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!

Issue 326: improved support for case sensitive (mixed case) identifiers without quotes when using DATABASE_TO_UPPER=FALSE.

The optimization for "group by" was not working correctly if the group by column was aliased in the select list.

--no commit message

Issue 351: MySQL mode: can not create a table with the column "KEY", and can not open databases where such a table already exists. (KEY is _not_ a keyword.)

Issue 351: MySQL mode: can not create a table with the column "KEY", and can not open databases where such a table already exists. (KEY is now a keyword.)