[
https://issues.apache.org/jira/browse/SPARK-59865?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Uroš Bojanić resolved SPARK-59865.
----------------------------------
Fix Version/s: 4.4.0
Resolution: Fixed
Issue resolved by pull request 59140
[https://github.com/apache/spark/pull/59140]
> Decide Oracle TimestampNTZ wall-clock conversion from JDBCOptions instead of
> column metadata
> --------------------------------------------------------------------------------------------
>
> Key: SPARK-59865
> URL: https://issues.apache.org/jira/browse/SPARK-59865
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 4.4.0
> Reporter: Aleksandar Trajković
> Assignee: Aleksandar Trajković
> Priority: Major
> Labels: pull-request-available
> Fix For: 4.4.0
>
>
> SPARK-58876 decides per column whether Oracle TimestampNTZ values are read
> and written as
> wall-clock LocalDateTime: OracleDialect stamps a marker into the column
> metadata, and writes
> rely on a new JdbcDialect.updateExtraColumnMetaForWrite hook. The decision is
> not per column;
> it depends only on the dialect and
> spark.sql.legacy.oracle.timestampNTZMapping.enabled.
> Replace the markers with a JdbcDialect.timestampNTZAsWallClock(options:
> JDBCOptions) hook, and
> snapshot the legacy flag in JDBCOptions so a flag change after planning does
> not change the
> planned conversion (the V2 JDBC catalog pins the snapshot into the table
> options).
> updateExtraColumnMetaForWrite is removed; it has not been released.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
