[
https://issues.apache.org/jira/browse/SPARK-4009?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sean Owen resolved SPARK-4009.
------------------------------
Resolution: Won't Fix
PR discussion says that this change should not be made.
> HiveTableScan should use makeRDDForTable instead of
> makeRDDForPartitionedTable for partitioned table when partitionPruningPred is
> None
> --------------------------------------------------------------------------------------------------------------------------------------
>
> Key: SPARK-4009
> URL: https://issues.apache.org/jira/browse/SPARK-4009
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Reporter: YanTang Zhai
>
> HiveTableScan should use makeRDDForTable instead of
> makeRDDForPartitionedTable for partitioned table when partitionPruningPred is
> None.
> If a table has many partitions for example more than 20 thousands while it
> has a few data for example less than 512MB, some sql querying the table will
> produce more than 20000 RDDs. The job would submit failed with exception:
> java stack overflow.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
