[
https://issues.apache.org/jira/browse/HUDI-7585?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ethan Guo updated HUDI-7585:
----------------------------
Description: The table schema resolver needs to read schema from the data
files (base or log files) to see whether _hoodie_operation field is present for
Flink CDC use cases. This can cause overhead of reading data file footers
multiple times. We should see if we can store a table config to indicate if
_hoodie_operation field is present in the table, or simplify the Flink CDC
format in Hudi 1.0 (thus no need of _hoodie_operation field and schema
resolver). (was: The table schema resolver needs to read schema from the data
files (base or log files) to see whether _hoodie_operation field is present for
Flink CDC use cases. This can cause overhead of reading data file footers
multiple times. We should see if we can store a table config to indicate if or
simplify the Flink CDC format in Hudi 1.0 (thus no need of _hoodie_operation
field and schema resolver).)
> Avoid reading log files for resolving schema for _hoodie_operation field
> ------------------------------------------------------------------------
>
> Key: HUDI-7585
> URL: https://issues.apache.org/jira/browse/HUDI-7585
> Project: Apache Hudi
> Issue Type: Improvement
> Reporter: Ethan Guo
> Assignee: Jing Zhang
> Priority: Major
> Fix For: 1.0.0
>
>
> The table schema resolver needs to read schema from the data files (base or
> log files) to see whether _hoodie_operation field is present for Flink CDC
> use cases. This can cause overhead of reading data file footers multiple
> times. We should see if we can store a table config to indicate if
> _hoodie_operation field is present in the table, or simplify the Flink CDC
> format in Hudi 1.0 (thus no need of _hoodie_operation field and schema
> resolver).
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
