[
https://issues.apache.org/jira/browse/FLINK-5864?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15887700#comment-15887700
]
ASF GitHub Bot commented on FLINK-5864:
---------------------------------------
Github user kl0u commented on the issue:
https://github.com/apache/flink/pull/3390
I agree that the solution is temporary but the new version (with the richer
pattern support) could take some time. As for the sorting, we could have a
"pre-traversal" of the transitions, as you had in your previous version of this
PR.
It could be a separate method that copies the state transitions list and
sorts it on the `action` field (make the `StateTransition` comparable) before
the for-loop in the `computeNextStates`.
This is not the most efficient solution, as it creates copies and sorts
every time, but it is enough until we update the whole `NFA`/`NFACompiler`.
> CEP: fix duplicate output patterns problem.
> -------------------------------------------
>
> Key: FLINK-5864
> URL: https://issues.apache.org/jira/browse/FLINK-5864
> Project: Flink
> Issue Type: Bug
> Components: CEP
> Affects Versions: 1.3.0
> Reporter: Kostas Kloudas
> Assignee: Dawid Wysakowicz
> Fix For: 1.3.0
>
>
> Currently when searching for a pattern {{a,b,c}} and we have input elements
> {{a -> b1 -> b2 ->c}} where {{b1}} and {{b2}} are both valid elements for the
> position b, then instead of having an output of 2 matched patterns: {{a, b1,
> c}} and {{a, b2, c}}, we have 4, with 2 copies of each valid pattern.
> The problem is with the creation of Dewey number, cause it is not increased
> on graph branching.
>
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
