rahil-c opened a new issue, #20204:
URL: https://github.com/apache/hudi/issues/20204
If the first completed commit of a table carries no schema, every later
commit on a record-index table fails, so the table can never be written to
again.
This happens when the first commit is empty and written with a null writer
schema, for example a Hudi Streamer sync whose source returned no data.
CommitUtils stores the schema as an empty string, nothing else records one, and
the record index update for the next data commit calls
HoodieTableMetadataUtil#tryResolveSchemaForTable, which throws:
HoodieException: Failed to generate RLI records for metadata table
Caused by: HoodieException: Failed to get latest columns for <basePath>
Caused by: HoodieSchemaNotFoundException: No schema found for table at
<basePath>
tryResolveSchemaForTable already returns Option.empty() when the table has
no completed commit, and its callers handle that. A completed commit without a
schema is the same situation and should resolve the same way instead of failing
the write.
--
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]
