hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Navis Ryu" <navis....@nexr.com>
Subject Re: Review Request: HIVE-2332 If all of the parameters of distinct functions are exists in group by columns, query fails in runtime
Date Wed, 07 Dec 2011 07:01:30 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/1314/
-----------------------------------------------------------

(Updated 2011-12-07 07:01:30.547976)


Review request for hive, John Sichi and Carl Steinbach.


Changes
-------

Adding null keys induced confusions especially for optimizers. This patch just modifies key-order
minimizing side effects.


Summary
-------

If all of the distinct params are in group by keys, union column reserved for distinct params
would not be added, which making problems initializing RS operator.

This patch is just a simple bypass adding dummy expression for the union column. Someone would
know better way to resolve the problem.


This addresses bug HIVE-2332.
    https://issues.apache.org/jira/browse/HIVE-2332


Diffs (updated)
-----

  ql/src/java/org/apache/hadoop/hive/ql/exec/ReduceSinkOperator.java 732a5aa 
  ql/src/java/org/apache/hadoop/hive/ql/plan/PlanUtils.java c6ae55d 
  ql/src/test/queries/clientpositive/groupby_distinct_samekey.q PRE-CREATION 
  ql/src/test/results/clientpositive/groupby_distinct_samekey.q.out PRE-CREATION 

Diff: https://reviews.apache.org/r/1314/diff


Testing
-------

added clientpositive/groupby_distinct_samekey.q


Thanks,

Navis


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message