You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@flink.apache.org by "Stephan Ewen (JIRA)" <ji...@apache.org> on 2014/12/14 15:59:13 UTC

[jira] [Created] (FLINK-1324) Iterations may fail when a cached data set based on a sort is not fully consumed

Stephan Ewen created FLINK-1324:
-----------------------------------

             Summary: Iterations may fail when a cached data set based on a sort is not fully consumed
                 Key: FLINK-1324
                 URL: https://issues.apache.org/jira/browse/FLINK-1324
             Project: Flink
          Issue Type: Bug
          Components: Local Runtime
    Affects Versions: 0.8-incubating
            Reporter: Stephan Ewen
            Assignee: Stephan Ewen
             Fix For: 0.8-incubating


This happens for example when a sort feeds into a merge join and the merge join exits early (zig zag merge determined that it can stop).

You can reproduce this bug by running the transitive closure in Scala (or Java with an enforced merge join) using the following toy data set.

{code}
Path(2, 1),
Path(4, 1),
Path(6, 3),
Path(8, 3),
Path(10, 1),
Path(12, 1),
Path(14, 3),
Path(16, 3),
Path(18, 1),
Path(20, 1) );
{code}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)