[
https://issues.apache.org/jira/browse/DRILL-4083?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15004880#comment-15004880
]
ASF GitHub Bot commented on DRILL-4083:
---------------------------------------
Github user jinfengni commented on the pull request:
https://github.com/apache/drill/pull/253#issuecomment-156584940
Overall LGTM.
In case that only constants are used and no column is required from scan
operator, the scan operator still has to go through the data source; the cpu
costing should not be 0.
+1
> Drill not using HiveDrillNativeParquetScan if no column is needed to be read
> from HIVE
> --------------------------------------------------------------------------------------
>
> Key: DRILL-4083
> URL: https://issues.apache.org/jira/browse/DRILL-4083
> Project: Apache Drill
> Issue Type: Bug
> Components: Storage - Hive
> Reporter: Sean Hsuan-Yi Chu
> Assignee: Sean Hsuan-Yi Chu
>
> For example, for a query such as:
> {code}
> Select count(*) from hive.parquetTable
> {code}
> would not use HiveDrillNativeParquetScan. However, the following query will
> use:
> {code}
> Select count(*) from hive.parquetTable where column > 0
> {code}
> Ideally, both should use the same HiveDrillNativeParquetScan
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
