You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "vinoyang (JIRA)" <ji...@apache.org> on 2019/02/18 02:18:00 UTC
[jira] [Commented] (FLINK-11639) Provide readSequenceFile for
Hadoop new API
[ https://issues.apache.org/jira/browse/FLINK-11639?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16770733#comment-16770733 ]
vinoyang commented on FLINK-11639:
----------------------------------
[~fhueske] What do you think about this issue? I am currently using Flink's batch API, but {{HadoopInputs}} only provides APIs for older version of {{SequenceFileInputFormat}}, and no new API for the new version of {{SequenceFileInputFormat}} has been found. If we introduce a new API, it helps to ensure symmetry, which seems to be more user-friendly, since it has the same method name and parameter list as {{readSequenceFile}}, then there will be conflicts. So, I recommend naming it {{readMapReduceSequenceFile}}. If such a name can cause confusion for the user, then we can also rename the original {{readSequenceFile}} to {{readMapRedSequenceFile}}.
> Provide readSequenceFile for Hadoop new API
> -------------------------------------------
>
> Key: FLINK-11639
> URL: https://issues.apache.org/jira/browse/FLINK-11639
> Project: Flink
> Issue Type: Improvement
> Components: Batch Connectors and Input/Output Formats
> Reporter: vinoyang
> Assignee: vinoyang
> Priority: Major
>
> Now {{HadoopInputs}} just provide a {{readSequenceFile}} for {{org.apache.hadoop.mapred.SequenceFileInputFormat}} , it would be better to provide another {{readSequenceFile}} for {{org.apache.hadoop.mapreduce.lib.input.SequenceFileInputFormat}}.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)