[jira] [Created] (FLINK-7666) ContinuousFileReaderOperator swallows chained watermarks

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

[jira] [Created] (FLINK-7666) ContinuousFileReaderOperator swallows chained watermarks

Shang Yuanchun (Jira)
Ufuk Celebi created FLINK-7666:
----------------------------------

             Summary: ContinuousFileReaderOperator swallows chained watermarks
                 Key: FLINK-7666
                 URL: https://issues.apache.org/jira/browse/FLINK-7666
             Project: Flink
          Issue Type: Improvement
          Components: Streaming Connectors
    Affects Versions: 1.3.2
            Reporter: Ufuk Celebi


I use event time and read from a (finite) file. I assign watermarks right after the {{ContinuousFileReaderOperator}} with parallelism 1.

{code}
env
  .readFile(new TextInputFormat(...), ...)
  .setParallelism(1)
  .assignTimestampsAndWatermarks(...)
  .setParallelism(1)
  .map()...
{code}

The watermarks I assign never progress through the pipeline.

I can work around this by inserting a {{shuffle()}} after the file reader or starting a new chain at the assigner:
{code}
env
  .readFile(new TextInputFormat(...), ...)
  .setParallelism(1)
  .shuffle()
  .assignTimestampsAndWatermarks(...)
  .setParallelism(1)
  .map()...
{code}



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)