ShengJun Zheng created KYLIN-4919:
-------------------------------------
Summary: Use NonShardingReadRDD/ShardingReadRdd automatically for
different queries
Key: KYLIN-4919
URL: https://issues.apache.org/jira/browse/KYLIN-4919
Project: Kylin
Issue Type: Improvement
Reporter: ShengJun Zheng
FilePruner will prune files in a segment cuboid dir by shard spec，then
create a ShardingReadRdd by default (to optimize join performance).
For queries without join, there is no need to create a shard RDD, it result
in unnecessary empty spark tasks. It's better to create a NonShardingReadRDD
after shard pruning.
Suggest that KYLIN should decide to use NonShardingReadRDD or
ShardingReadRdd automatically for different queries.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
