hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From kelly zhang <liyun.zh...@intel.com>
Subject Re: Review Request 61839: HIVE-16823:ArrayIndexOutOfBoundsException" in spark_vectorized_dynamic_partition_pruning.q
Date Wed, 23 Aug 2017 08:33:55 GMT

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/61839/#review183579
-----------------------------------------------------------




ql/src/test/results/clientpositive/spark/spark_vectorized_dynamic_partition_pruning.q.out
Line 17 (original)
<https://reviews.apache.org/r/61839/#comment259630>

    the changes is because we have not add "-- SORT_QUERY_RESULTS"



ql/src/test/results/clientpositive/spark/spark_vectorized_dynamic_partition_pruning.q.out
Line 875 (original)
<https://reviews.apache.org/r/61839/#comment259631>

    The reason why Map5 and Map6 is combined is because HIVE-11297



ql/src/test/results/clientpositive/spark/spark_vectorized_dynamic_partition_pruning.q.out
Line 2337 (original)
<https://reviews.apache.org/r/61839/#comment259632>

    Why Map5 and Map6 is combined is because HIVE-11297



ql/src/test/results/clientpositive/spark/spark_vectorized_dynamic_partition_pruning.q.out
Lines 2891-2918 (patched)
<https://reviews.apache.org/r/61839/#comment259633>

    Map 8 is added. Map 8 is necessary because in query 
    {code}
    
     select count(*) from srcpart join srcpart_date on (srcpart.ds = srcpart_date.ds) join
srcpart_hour on (srcpart.hr = srcpart_hour.hr) 
    {code}
    
    srcpart_date and srcpart_hour are all small tables, both of them should be converted to
dpp operator tree because srcpart has two partitions "srcpart.ds" and "srcpart.hr"



ql/src/test/results/clientpositive/spark/spark_vectorized_dynamic_partition_pruning.q.out
Line 5331 (original)
<https://reviews.apache.org/r/61839/#comment259637>

    will investigate why Map4 is removed in Line 5312 in query 
    " EXPLAIN select count(*) from srcpart full outer join srcpart_date on (srcpart.ds = srcpart_date.ds)
where srcpart_date.`date` = '2008-04-08'"


- kelly zhang


On Aug. 23, 2017, 7:29 a.m., kelly zhang wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/61839/
> -----------------------------------------------------------
> 
> (Updated Aug. 23, 2017, 7:29 a.m.)
> 
> 
> Review request for hive, Rui Li and Sahil Takiar.
> 
> 
> Bugs: HIVE-16823
>     https://issues.apache.org/jira/browse/HIVE-16823
> 
> 
> Repository: hive-git
> 
> 
> Description
> -------
> 
> More detailed see jira board
> 
> 
> Diffs
> -----
> 
>   ql/src/java/org/apache/hadoop/hive/ql/parse/spark/SparkCompiler.java 73e596e 
>   ql/src/test/results/clientpositive/spark/spark_dynamic_partition_pruning.q.out 63a9548

>   ql/src/test/results/clientpositive/spark/spark_dynamic_partition_pruning_2.q.out cc7819c

>   ql/src/test/results/clientpositive/spark/spark_dynamic_partition_pruning_3.q.out 82e746a

>   ql/src/test/results/clientpositive/spark/spark_dynamic_partition_pruning_mapjoin_only.q.out
41e98a1 
>   ql/src/test/results/clientpositive/spark/spark_vectorized_dynamic_partition_pruning.q.out
699fcc6 
> 
> 
> Diff: https://reviews.apache.org/r/61839/diff/1/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> kelly zhang
> 
>


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