mxm commented on code in PR #17018:
URL: https://github.com/apache/iceberg/pull/17018#discussion_r3522618528
##########
flink/v2.1/flink/src/main/java/org/apache/iceberg/flink/maintenance/operator/EqualityConvertPKIndex.java:
##########
@@ -250,14 +261,22 @@ private void resolveDeletes(Collector<DVPosition> out)
throws Exception {
// indexed sequence is not comparable to the staging delete's; event-time
ordering already
// prevents false deletion.
Long deleteSeq = resolveSequenceNumber.value();
+ // A partitioned delete applies only to data rows of the same spec. An
unpartitioned delete
+ // (GLOBAL_DELETE_SPEC_ID) applies to every spec. deleteSpecIds holds
every scope seen this
+ // cycle (usually one).
+ List<Integer> deleteSpecIds = Lists.newArrayList(resolveSpecIds.get());
Review Comment:
I was considering this as well. The actual list of active spec ids will
likely be relatively small and the cost of iterating through them marginal. Set
may be more expensive here, but it probably does not matter much. Updated.
--
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]
