Microsoft SQL Server
CURRENT_DATE and CURRENT_TIME are now translated to CONVERT(DATE, GETDATE()) and CONVERT(TIME, GETDATE()) instead of plain GETDATE(), so the results are DATE/TIME typed and comparisons against datetime columns may behave differently. EXTRACT uses DATEPART and JPQL time literals are rendered as CAST('hh:mm:ss' AS TIME).
To restore the previous SQL set openjpa.jdbc.DBDictionary=sqlserver(CurrentDateFunction=GETDATE(),CurrentTimeFunction=GETDATE()).
SQLServerDictionary declares no NextSequenceQuery, although SQL Server has had sequence objects since 2012, so a named @SequenceGenerator that used to fail is now backed by a counter row instead of by the sequence it names.
To draw from that sequence, configure the query - see @SequenceGenerator on a database without sequences.
EXTRACT(WEEK FROM …) now returns the ISO week, as the specification defines it, by rendering DATEPART(ISO_WEEK, x) (OPENJPA-3064).
Before, SQL Server counted the weeks from January 1st, so queries on such dates return different values: for example 2023-12-31 is now week 52 instead of 53.
EXTRACT(SECOND FROM …) now includes the fraction of the second, as the specification requires, by adding DATEPART(NANOSECOND, x) / 1000000000.0 to the whole seconds that DATEPART(SECOND, x) returns.
When the schema tool drops a single column, for example with SchemaAction='drop' on a table that holds columns of classes that are not mapped, it now drops the primary key, unique constraints, indexes, default constraints and foreign keys of the table that contain the column first, as other databases do with the column itself; SQL Server refused to drop such a column before.
The objects and the column are dropped in one transaction: if the column cannot be dropped, for example because a foreign key of another table references the primary key, nothing is dropped.