hive-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Zoltan Haindrich (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HIVE-18225) disaggreement in results; vector_decimal_mapjoin.q
Date Tue, 05 Dec 2017 17:43:00 GMT

    [ https://issues.apache.org/jira/browse/HIVE-18225?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16278931#comment-16278931
] 

Zoltan Haindrich commented on HIVE-18225:
-----------------------------------------

[~sershe], [~teddy.choi]: I'm not sure wheteher this is connect to the other issue; but I've
to say that I see the llap_fast related result diff almost all the time: when  {{hive.stats.column.autogather}}
to true...this might be possibly a more stable result difference...

> disaggreement in results; vector_decimal_mapjoin.q
> --------------------------------------------------
>
>                 Key: HIVE-18225
>                 URL: https://issues.apache.org/jira/browse/HIVE-18225
>             Project: Hive
>          Issue Type: Bug
>            Reporter: Zoltan Haindrich
>
> one of these out files have changed for me...and it turned out that it's not clear which
output should be the expected; since we have different version recorded :D
> for the last query in {{vector_decimal_mapjoin.q}} ; spark and clidriver gives 1 result;
in the meanwhile llap returns a whole bunch of results.
> {code}
> $ tail `find ql -name vector_decimal_mapjoin.q.out`
> ==> ql/src/test/results/clientpositive/vector_decimal_mapjoin.q.out <==
> PREHOOK: type: QUERY
> PREHOOK: Input: default@t1_small
> PREHOOK: Input: default@t2_small
> #### A masked pattern was here ####
> POSTHOOK: query: select t1_small.`dec`, t1_small.value_dec, t2_small.`dec`, t2_small.value_dec
from t1_small join t2_small on (t1_small.`dec`=t2_small.`dec`)
> POSTHOOK: type: QUERY
> POSTHOOK: Input: default@t1_small
> POSTHOOK: Input: default@t2_small
> #### A masked pattern was here ####
> 89.00	15.09	89	15
> ==> ql/src/test/results/clientpositive/llap/vector_decimal_mapjoin.q.out <==
> 9.00	48.96	9	34
> 9.00	48.96	9	38
> 9.00	48.96	9	41
> 9.00	48.96	9	42
> 9.00	48.96	9	45
> 9.00	48.96	9	48
> 9.00	48.96	9	49
> 9.00	48.96	9	5
> 9.00	48.96	9	7
> 9.00	48.96	9	7
> ==> ql/src/test/results/clientpositive/spark/vector_decimal_mapjoin.q.out <==
> PREHOOK: type: QUERY
> PREHOOK: Input: default@t1_small
> PREHOOK: Input: default@t2_small
> #### A masked pattern was here ####
> POSTHOOK: query: select t1_small.`dec`, t1_small.value_dec, t2_small.`dec`, t2_small.value_dec
from t1_small join t2_small on (t1_small.`dec`=t2_small.`dec`)
> POSTHOOK: type: QUERY
> POSTHOOK: Input: default@t1_small
> POSTHOOK: Input: default@t2_small
> #### A masked pattern was here ####
> 89.00	15.09	89	15
> {code}
> I think the TestClidriver based one can be "changed" to the other result; by setting
{{hive.stats.column.autogather=true}}.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message