[
https://issues.apache.org/jira/browse/BEAM-4587?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16517208#comment-16517208
]
Łukasz Gajowy commented on BEAM-4587:
-------------------------------------
I tested the scenarios above - no issues found, the IO works correctly.
Here's the code with README.md containing instructions I used:
[https://github.com/lgajowy/beam/tree/spark-flink-parquet-example] - the beam
part in /examples module (ParquetIO is not officially released yet, so it was
the easiest way)
[https://github.com/lgajowy/laying-parquet] - spark & flink code to do the job.
> Test interoperability between Spark, Flink and Beam in terms of
> reading/writing Parquet files
> ---------------------------------------------------------------------------------------------
>
> Key: BEAM-4587
> URL: https://issues.apache.org/jira/browse/BEAM-4587
> Project: Beam
> Issue Type: Task
> Components: io-java-parquet
> Reporter: Łukasz Gajowy
> Assignee: Łukasz Gajowy
> Priority: Minor
>
> Since ParquetIO is merged to master, we should test how it behaves with
> parquet files created by native Spark and Flink applications.
> More specifically, we should:
> - test if files created by Flink/Spark can be read successfully using
> ParquetIO in Beam
> - test if files created by beam using ParquetIO in Beam can be read using
> Flink/Spark native application.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
