[
https://issues.apache.org/jira/browse/HIVE-25772?focusedWorklogId=690037&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-690037
]
ASF GitHub Bot logged work on HIVE-25772:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 03/Dec/21 14:19
Start Date: 03/Dec/21 14:19
Worklog Time Spent: 10m
Work Description: szlta opened a new pull request #2842:
URL: https://github.com/apache/hive/pull/2842
Currently Hive relies on PartitionedFanoutWriter to write records to Iceberg
tables. This has a big disadvantage when it comes to writing many-many
partitions, as it keeps a file handle open to each of the partitions.
For some file systems like S3, there can be a big resource waste due to
keeping bytebuffers for each of the handles, that may result in OOM scenarios.
ClusteredWriter will only write one file at a time, and will not keep other
file handles open, but it will expect that the input data is sorted by
partition keys.
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
Issue Time Tracking
-------------------
Worklog Id: (was: 690037)
Remaining Estimate: 0h
Time Spent: 10m
> Use ClusteredWriter when writing to Iceberg tables
> --------------------------------------------------
>
> Key: HIVE-25772
> URL: https://issues.apache.org/jira/browse/HIVE-25772
> Project: Hive
> Issue Type: Improvement
> Reporter: Ádám Szita
> Assignee: Ádám Szita
> Priority: Major
> Labels: pull-request-available
> Time Spent: 10m
> Remaining Estimate: 0h
>
> Currently Hive relies on PartitionedFanoutWriter to write records to Iceberg
> tables. This has a big disadvantage when it comes to writing many-many
> partitions, as it keeps a file handle open to each of the partitions.
> For some file systems like S3, there can be a big resource waste due to
> keeping bytebuffers for each of the handles, that may result in OOM scenarios.
> ClusteredWriter will only write one file at a time, and will not keep other
> file handles open, but it will expect that the input data is sorted by
> partition keys.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
