[jira] [Commented] (HIVE-23043) Hive TPCDS: mr and tez occur different result about query15

2020-03-18 Thread zhangbutao (Jira)


[ 
https://issues.apache.org/jira/browse/HIVE-23043?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17061619#comment-17061619
 ] 

zhangbutao commented on HIVE-23043:
---

In addition, 2 GB orc data (tpcds_bin_partitioned_orc_2), tez get the same 
result  as mr does. So, it seems that result could be the same with small data.

[~ngangam] [~hashutosh] ,do you have any suggestions? Thanks.

> Hive TPCDS: mr and tez occur different result about query15
> ---
>
> Key: HIVE-23043
> URL: https://issues.apache.org/jira/browse/HIVE-23043
> Project: Hive
>  Issue Type: Bug
>Affects Versions: 3.1.0, 3.1.1
>Reporter: zhangbutao
>Priority: Major
> Attachments: explain_mr_query15.text, explain_tez_query15.text, 
> mr_query15_result.text, query15.sql, tez_query15_result.text
>
>
> Test info: hive-testbench, Hive TPCDS, 1TB orc data 
> (tpcds_bin_partitioned_orc_1000).
> mr and tez occur different result about query15.  



--
This message was sent by Atlassian Jira
(v8.3.4#803005)


[jira] [Commented] (HIVE-23043) Hive TPCDS: mr and tez occur different result about query15

2020-03-18 Thread zhangbutao (Jira)


[ 
https://issues.apache.org/jira/browse/HIVE-23043?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17061574#comment-17061574
 ] 

zhangbutao commented on HIVE-23043:
---

Spark SQL has the same result with tez.  So tez might be right and mr is wrong. 
 Can anyone help explain the question?  Thanks!

> Hive TPCDS: mr and tez occur different result about query15
> ---
>
> Key: HIVE-23043
> URL: https://issues.apache.org/jira/browse/HIVE-23043
> Project: Hive
>  Issue Type: Bug
>Affects Versions: 3.1.0, 3.1.1
>Reporter: zhangbutao
>Priority: Major
> Attachments: mr_query15_result.text, tez_query15_result.text
>
>
> Test info: hive-testbench, Hive TPCDS, 1TB orc data 
> (tpcds_bin_partitioned_orc_1000).
> mr and tez occur different result about query15.  



--
This message was sent by Atlassian Jira
(v8.3.4#803005)