You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@hive.apache.org by "Denys Kuzmenko (Jira)" <ji...@apache.org> on 2023/07/21 07:35:00 UTC

[jira] [Commented] (HIVE-27437) Vectorization: VectorizedOrcRecordReader does not reset VectorizedRowBatch after processing

    [ https://issues.apache.org/jira/browse/HIVE-27437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17745470#comment-17745470 ] 

Denys Kuzmenko commented on HIVE-27437:
---------------------------------------

Merged to master.
[~maswin], thank you for the contribution!

> Vectorization: VectorizedOrcRecordReader does not reset VectorizedRowBatch after processing
> -------------------------------------------------------------------------------------------
>
>                 Key: HIVE-27437
>                 URL: https://issues.apache.org/jira/browse/HIVE-27437
>             Project: Hive
>          Issue Type: Task
>            Reporter: Alagappan Maruthappan
>            Assignee: Alagappan Maruthappan
>            Priority: Major
>              Labels: pull-request-available
>
> There seems to be a memory leak in VectorizedOrcRecordReader. When 
> MapColumnVector or ListColumnVector is used and the VectorizedRowBatch is not reset after every read, the vector keeps growing and spending a lot of time assigning memory. 
> The reset happens in VectorizedParquetRecordReader -
> https://github.com/apache/hive/blob/f78ca5df80c0bcb566f0915cda65112268df492c/ql/src/java/org/apache/hadoop/hive/ql/io/parquet/vector/VectorizedParquetRecordReader.java#L400



--
This message was sent by Atlassian Jira
(v8.20.10#820010)