[ https://issues.apache.org/jira/browse/HIVE-15682?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15863880#comment-15863880 ]
Xuefu Zhang commented on HIVE-15682: ------------------------------------ Correct. It would be very helpful if you can try the following queries: 1. w/o HIVE-15580, w/ HIVE-15580, and w/ HIVE-15580+HIVE-15682 for an order by query like: {code} select count(*) from (select request_lat from dwh.fact_trip where datestr > '2017-01-27' order by request_lat) x; {code} 2. w/o HIVE-15580, w/ HIVE-15580 for an group by query like: {code} select count(*) from (select driver_uuid, avg(base_fare_usd) from dwh.fact_trip where datestr > '2017-01-01' group by driver_uuid) x; {code} Also, it would be great if you can also analyze the benchmark result, especially confirming why HIVE-15682 has adverse performance impact. Thanks. > Eliminate per-row based dummy iterator creation > ----------------------------------------------- > > Key: HIVE-15682 > URL: https://issues.apache.org/jira/browse/HIVE-15682 > Project: Hive > Issue Type: Improvement > Components: Spark > Affects Versions: 2.2.0 > Reporter: Xuefu Zhang > Assignee: Xuefu Zhang > Fix For: 2.2.0 > > Attachments: HIVE-15682.patch > > > HIVE-15580 introduced a dummy iterator per input row which can be eliminated. > This is because {{SparkReduceRecordHandler}} is able to handle single key > value pairs. We can refactor this part of code 1. to remove the need for a > iterator and 2. to optimize the code path for per (key, value) based (instead > of (key, value iterator)) processing. It would be also great if we can > measure the performance after the optimizations and compare to performance > prior to HIVE-15580. -- This message was sent by Atlassian JIRA (v6.3.15#6346)