flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Fabian Hueske (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-12008) Support read a whole directory or multiple input data files for read apis of HadoopInputs
Date Mon, 08 Apr 2019 09:27:00 GMT

    [ https://issues.apache.org/jira/browse/FLINK-12008?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16812265#comment-16812265
] 

Fabian Hueske commented on FLINK-12008:
---------------------------------------

Yes, IIRC we recently added support for multiple paths to Flink's {{FileInputFormat}} as well.

> Support read a whole directory or multiple input data files for read apis of HadoopInputs
> -----------------------------------------------------------------------------------------
>
>                 Key: FLINK-12008
>                 URL: https://issues.apache.org/jira/browse/FLINK-12008
>             Project: Flink
>          Issue Type: Improvement
>          Components: Connectors / Hadoop Compatibility
>            Reporter: vinoyang
>            Assignee: vinoyang
>            Priority: Major
>
> Currently, the read APIs provided by {{HadoopInputs}} only can read one path. I think
it's not strong enough. We should support read a whole directory or multiple input files.
> Hadoop provides {{org.apache.hadoop.mapred.FileInputFormat.setInputPaths()}} to support
this requirement. 
> Spark's {{sequenceFile}} API calls this API([https://github.com/apache/spark/blob/master/core/src/main/scala/org/apache/spark/SparkContext.scala#L1049].)
> Flink calls {{org.apache.hadoop.mapred.FileInputFormat.addInputPath}} which only  supports
one path.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Mime
View raw message