yihua opened a new issue, #925:
URL: https://github.com/apache/hudi-rs/issues/925
With a non-global index (the default SIMPLE), a table written by Java can
store the same record key in two partitions. `table.delete("id = 'k'")` takes
the record-key path, which builds keys with an empty partition, and
`tag_location` resolves each key to a single stored copy, so only one
partition's row is deleted, on both copy-on-write and merge-on-read. A delete
on a non-key column that matches the same rows removes both, so the same rows
get different results depending on which column the predicate names.
A key-only delete should remove every stored copy of the key. Found while
reviewing #851, which makes the copy it picks deterministic.
--
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]
