Jonathan Vexler created HUDI-8166:
-------------------------------------
Summary: Calculate "mandatory partition fields to read from the
file" inside of the filegroup reader so we can have cross engine support
Key: HUDI-8166
URL: https://issues.apache.org/jira/browse/HUDI-8166
Project: Apache Hudi
Issue Type: Improvement
Components: reader-core
Reporter: Jonathan Vexler
[https://github.com/apache/hudi/pull/11770] [Hudi-5807] adds support to spark
fg reader to read the actual values from timestamp keygen columns. We can
probably pull that logic into the fg reader. For hive the logic will be
different though. Because partition path has to be constant so we create an
artificial field iirc. And also we might need to pass in the partition value
into the fg reader for appending
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
