mxm opened a new pull request, #17018:
URL: https://github.com/apache/iceberg/pull/17018

   The PK index keys rows by (spec id, equality values), so a delete only 
matches data written under the same spec. However, according to the Iceberg 
table spec, while a partitioned delete applies only within its own spec, an 
unpartitioned equality delete must apply across every spec.
   
   This change drops the spec id from the key, so all rows with the same 
equality values co-locate on one shard. We scope at resolve time instead: 
RESOLVE_DELETE carries the delete's spec id, or a GLOBAL_DELETE_SPEC_ID when 
the delete is unpartitioned. The index deletes a row only when the delete is 
global or the row's spec id matches. DVPosition already carries each row's spec 
id, so no extra row state is needed.


-- 
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]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to