[
https://issues.apache.org/jira/browse/DRILL-456?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13968628#comment-13968628
]
Aman Sinha commented on DRILL-456:
----------------------------------
I believe this is just a bug in streaming aggregate outputting the first row
thinking that it is not the same as a 'previous' row's grouping key. We just
need to keep track of the first row. I have a potential fix that I am testing.
> Inconsistent results between Hive and Drill (an extra row appeared in Drill)
> ----------------------------------------------------------------------------
>
> Key: DRILL-456
> URL: https://issues.apache.org/jira/browse/DRILL-456
> Project: Apache Drill
> Issue Type: Bug
> Reporter: Neeraja
> Assignee: Aman Sinha
>
> Tried a simple query as below.
> The results look different between Hive and Drill (note the extra result row
> in case of Drill)
> Query:
> select l_returnflag,l_linestatus, count(*) from lineitem group by
> l_returnflag, l_linestatus order by l_returnflag,l_linestatus;
> Hive results:
> A F 147790
> N F 3765
> N O 300716
> R F 148301
> Drill results:
> +------------+--------------+--------------+
> | EXPR$2 | l_returnflag | l_linestatus |
> +------------+--------------+--------------+
> | 147790 | A | F |
> | 3765 | N | F |
> | 0 | N | O |
> | 300716 | N | O |
> | 148301 | R | F |
> +------------+--------------+--------------+
--
This message was sent by Atlassian JIRA
(v6.2#6252)
