[
https://issues.apache.org/jira/browse/SPARK-11881?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15018086#comment-15018086
]
Marius Van Niekerk commented on SPARK-11881:
--------------------------------------------
I have a PR for this incoming
> [SQL] JDBC Postgres fetchsize parameter ignored
> ------------------------------------------------
>
> Key: SPARK-11881
> URL: https://issues.apache.org/jira/browse/SPARK-11881
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 1.6.0
> Reporter: Marius Van Niekerk
>
> Reference:
> https://jdbc.postgresql.org/documentation/head/query.html#query-with-cursor
> In order for PostgreSQL to honor the fetchSize non-zero setting, its
> Connection.autoCommit needs to be set to false. Otherwise, it will just
> quietly ignore the fetchSize setting.
> Since the JDBC driver doesn’t serialize the result data in its memory, we
> don’t want to have 2-3 times of the memory created in the driver first before
> Spark can fetch anything into its serialized data storage.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
