[
https://issues.apache.org/jira/browse/SPARK-6593?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14385706#comment-14385706
]
Sean Owen commented on SPARK-6593:
----------------------------------
At this level though, what's a bad split? a line of text that doesn't parse as
expected? that's application-level logic. Given how little the framework knows,
this would amount to ignoring a partition if there was any error in computing
it, which seems too coarse to encourage people to use. You can of course handle
this in the application logic -- catch the error, return nothing, log it, add
to a counter, etc.
> Provide option for HadoopRDD to skip bad data splits.
> -----------------------------------------------------
>
> Key: SPARK-6593
> URL: https://issues.apache.org/jira/browse/SPARK-6593
> Project: Spark
> Issue Type: Improvement
> Components: Spark Core
> Affects Versions: 1.3.0
> Reporter: Dale Richardson
> Priority: Minor
>
> When reading a large amount of files from HDFS eg. with
> sc.textFile("hdfs:///user/cloudera/logs*.gz"). If a single split is corrupted
> then the entire job is canceled. As default behaviour this is probably for
> the best, but it would be nice in some circumstances where you know it will
> be ok to have the option to skip the corrupted portion and continue the job.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
