[
https://issues.apache.org/jira/browse/SPARK-12850?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15114415#comment-15114415
]
Xiao Li commented on SPARK-12850:
---------------------------------
Doing the design and prototype. Notice a few issues we need to consider:
- Partitioning columns could have overlapping columns with the bucketing
columns;
- The predicates we can use for bucket pruning: EqualTo, EqualNullSafe, IsNull,
In, InSet;
- Need to support mixed And and Or in the filters;
- After generating the bucket set we need to scan, we should remove the
corresponding filters, if possible;
Maybe, I will just submit a simplified version at first.
> Support bucket pruning (predicate pushdown for bucketed tables)
> ---------------------------------------------------------------
>
> Key: SPARK-12850
> URL: https://issues.apache.org/jira/browse/SPARK-12850
> Project: Spark
> Issue Type: Sub-task
> Components: SQL
> Reporter: Reynold Xin
>
> We now support bucketing. One optimization opportunity is to push some
> predicates into the scan to skip scanning files that definitely won't match
> the values.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
