[
https://issues.apache.org/jira/browse/FLINK-9593?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16524984#comment-16524984
]
ASF GitHub Bot commented on FLINK-9593:
---------------------------------------
Github user kl0u commented on a diff in the pull request:
https://github.com/apache/flink/pull/6171#discussion_r198474417
--- Diff:
flink-libraries/flink-cep/src/main/java/org/apache/flink/cep/nfa/NFA.java ---
@@ -330,77 +328,85 @@ private boolean isStateTimedOut(final
ComputationState state, final long timesta
}
}
- discardComputationStatesAccordingToStrategy(
- sharedBuffer, computationStates, result,
afterMatchSkipStrategy);
+ if (!potentialMatches.isEmpty()) {
+ nfaState.setStateChanged();
+ }
+
+ List<Map<String, List<T>>> result = new ArrayList<>();
+ if (afterMatchSkipStrategy.isSkipStrategy()) {
+ processMatchesAccordingToSkipStrategy(sharedBuffer,
+ nfaState,
+ afterMatchSkipStrategy,
+ potentialMatches,
+ result);
+ } else {
+ for (ComputationState match : potentialMatches) {
+
result.add(sharedBuffer.materializeMatch(sharedBuffer.extractPatterns(match.getPreviousBufferEntry(),
--- End diff --
Instead of accessing the state for every match, why not passing all the
matches to the shared buffer, and try to fetch the common ones only once. If 2
matches A and B share event with id = 2, we fetch from state only once.
> Unify AfterMatch semantics with SQL MATCH_RECOGNIZE
> ---------------------------------------------------
>
> Key: FLINK-9593
> URL: https://issues.apache.org/jira/browse/FLINK-9593
> Project: Flink
> Issue Type: Improvement
> Components: CEP
> Reporter: Dawid Wysakowicz
> Assignee: Dawid Wysakowicz
> Priority: Major
> Labels: pull-request-available
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
