You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@hudi.apache.org by "Raymond Xu (Jira)" <ji...@apache.org> on 2022/05/03 09:09:00 UTC
[jira] [Updated] (HUDI-3979) Make sure Hudi Relations are only fetching strictly required columns
[ https://issues.apache.org/jira/browse/HUDI-3979?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Raymond Xu updated HUDI-3979:
-----------------------------
Sprint: Hudi-Sprint-Apr-25, Hudi-Sprint-May-02 (was: Hudi-Sprint-Apr-25)
> Make sure Hudi Relations are only fetching strictly required columns
> --------------------------------------------------------------------
>
> Key: HUDI-3979
> URL: https://issues.apache.org/jira/browse/HUDI-3979
> Project: Apache Hudi
> Issue Type: Bug
> Reporter: Alexey Kudinkin
> Assignee: Alexey Kudinkin
> Priority: Blocker
> Labels: pull-request-available
> Fix For: 0.12.0
>
>
> In 0.11 we were able to considerably optimize data throughput for MOR tables in XXX: previously MOR table would read the whole table (full row with all columns) even for a simple `df.count()` query.
>
> This has been optimized to just fetch the _required_ columns for each table-type (COW/MOR). However, we can optimize this even further:
> # COW tables do not require _any columns at all_
> # MOR tables do required primary-key and pre-combine-key columns only when actual merging w/ updated records (from delta-log files) is performed. Otherwise, they could avoid reading even these cols.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)