hive-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF GitHub Bot (Jira)" <>
Subject [jira] [Work logged] (HIVE-24139) VectorGroupByOperator is not flushing hash table entries as needed
Date Thu, 10 Sep 2020 16:08:00 GMT


ASF GitHub Bot logged work on HIVE-24139:

                Author: ASF GitHub Bot
            Created on: 10/Sep/20 16:07
            Start Date: 10/Sep/20 16:07
    Worklog Time Spent: 10m 
      Work Description: mustafaiman commented on a change in pull request #1481:

File path: ql/src/java/org/apache/hadoop/hive/ql/exec/vector/
@@ -427,8 +428,12 @@ public void initialize(Configuration hconf) throws HiveException {
       LOG.debug("using hash aggregation processing mode");
+      reusableBufferSize = (int) (maxHtEntries * percentEntriesToFlush);
+      reusableAggregationBufferRows = new ArrayDeque<>(reusableBufferSize);
       if (keyWrappersBatch.getVectorHashKeyWrappers()[0] instanceof VectorHashKeyWrapperGeneral)

Review comment:
   The array that is returned from `keyWrappersBatch.getVectorHashKeyWrappers()` consists
of only one type of wrappers, isn't it? So the first element is representative of all the
elements in that array.
   We resort to this optimization only when the batch consists of VectorHashKeyWrapperGeneral.
The other wrapper types have only one two primitive types. If needed, only thing we need to
do is just override their copyKey(KeyWrapper) method and generalize this `instanceOf` check
to cover them too. Right now, I dont think that is very important.
   So when the first element is not VectorHashKeyWrapperGeneral, the rest of the array are
not either. In that case, we just skip reusing.

This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:

Issue Time Tracking

    Worklog Id:     (was: 481548)
    Time Spent: 0.5h  (was: 20m)

> VectorGroupByOperator is not flushing hash table entries as needed
> ------------------------------------------------------------------
>                 Key: HIVE-24139
>                 URL:
>             Project: Hive
>          Issue Type: Bug
>            Reporter: Mustafa Iman
>            Assignee: Mustafa Iman
>            Priority: Major
>              Labels: pull-request-available
>          Time Spent: 0.5h
>  Remaining Estimate: 0h
> After introduced a bug where copyKey
mutates some key wrappers while copying. This Jira is to fix it.

This message was sent by Atlassian Jira

View raw message