[
https://issues.apache.org/jira/browse/SPARK-11989?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Reynold Xin updated SPARK-11989:
--------------------------------
Assignee: Christian Kurz
> Spark JDBC write only works on techologies with transaction support
> -------------------------------------------------------------------
>
> Key: SPARK-11989
> URL: https://issues.apache.org/jira/browse/SPARK-11989
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 1.5.2
> Reporter: Christian Kurz
> Assignee: Christian Kurz
> Original Estimate: 4h
> Remaining Estimate: 4h
>
> Writing DataFrames out to a JDBC destination currently requires the JDBC
> driver/ database to support transaction. This is because
> spark/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/jdbc/JdbcUtils.scala
> *always* calls commit()/rollback().
> Some technologies do not support transactions and their drivers will throw an
> exception if commit()/rollback() is used. For these technologies (like
> Progress JDBC Driver for Cassandra) this is a blocking problem.
> Prior to using transaction support JdbcUtils.scala needs to check whether the
> drivers does support transaction. Check can be done via
> conn.getMetaData().supportsDataManipulationTransactionsOnly() ||
> conn.getMetaData().supportsDataDefinitionAndDataManipulationTransactions()
> The working code change can be seen
> [here|https://github.com/CK50/spark/blob/branch-1.6_non-transactional/sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/jdbc/JdbcUtils.scala]
> [Pull request|https://github.com/apache/spark/pull/9973]
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
