[
https://issues.apache.org/jira/browse/HIVE-14362?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15431939#comment-15431939
]
Gopal V commented on HIVE-14362:
--------------------------------
[~pxiong]: tested this patch - running explain analyze seems to disable
vectorization for all queries after that point.
{code}
+ HiveConf.setBoolVar(conf, HiveConf.ConfVars.HIVE_VECTORIZATION_ENABLED,
false);
{code}
And explain analyze does not actually work.
{code}
2016-08-23T01:13:10,961 INFO [667a4e5f-6194-438f-85d6-339aca3ebecc main]
physical.AnnotateRunTimeStatsOptimizer: setRuntimeStatsDir for RS_8
2016-08-23T01:13:10,962 INFO [667a4e5f-6194-438f-85d6-339aca3ebecc main]
fs.FSStatsPublisher: created :
file:/tmp/gopal/667a4e5f-6194-438f-85d6-339aca3ebecc/hive_2016-08-23_01-13-10_705_7555853843090786759-1/-local-10000/RS_8
{code}
The paths for output are in local dirs, not the HDFS dirs - so the stats
written on a machine are not making their way back to the HiveServer2 box.
{code}
Caused by: org.apache.hadoop.hive.ql.metadata.HiveException: [Error 30002]:
StatsPublisher cannot be connected to.There was a error while connecting to the
StatsPublisher, and retrying might help. If you dont want the query to fail
because accurate statistics could not be collected, set
hive.stats.reliable=false
at
org.apache.hadoop.hive.ql.exec.Operator.publishRunTimeStats(Operator.java:1444)
at org.apache.hadoop.hive.ql.exec.Operator.closeOp(Operator.java:723)
at
org.apache.hadoop.hive.ql.exec.TableScanOperator.closeOp(TableScanOperator.java:270)
at org.apache.hadoop.hive.ql.exec.Operator.close(Operator.java:691)
at org.apache.hadoop.hive.ql.exec.Operator.close(Operator.java:705)
at
org.apache.hadoop.hive.ql.exec.tez.MapRecordProcessor.close(MapRecordProcessor.java:433)
{code}
> Support explain analyze in Hive
> -------------------------------
>
> Key: HIVE-14362
> URL: https://issues.apache.org/jira/browse/HIVE-14362
> Project: Hive
> Issue Type: New Feature
> Reporter: Pengcheng Xiong
> Assignee: Pengcheng Xiong
> Attachments: HIVE-14362.01.patch, HIVE-14362.02.patch,
> compare_on_cluster.pdf
>
>
> Right now all the explain levels only support stats before query runs. We
> would like to have an explain analyze similar to Postgres for real stats
> after query runs. This will help to identify the major gap between
> estimated/real stats and make not only query optimization better but also
> query performance debugging easier.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)