[jira] [Comment Edited] (APEXMALHAR-1911) AbstractFileOutputOperator roll over boundary case

["Yogi Devendra (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/APEXMALHAR-1911?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15079377#comment-15079377 ] 

Yogi Devendra edited comment on APEXMALHAR-1911 at 3/7/16 11:51 AM:
--------------------------------------------------------------------

[~milindbee] Can you please add mohit to the users list. Currently, I could not add him as a watcher for Malhar JIRAs?


was (Author: devendra):
[~milind] Can you please add mohit to the users list. Currently, I could not add him as a watcher for Malhar JIRAs?

> AbstractFileOutputOperator roll over boundary case
> --------------------------------------------------
>
>                 Key: APEXMALHAR-1911
>                 URL: https://issues.apache.org/jira/browse/APEXMALHAR-1911
>             Project: Apache Apex Malhar
>          Issue Type: Bug
>            Reporter: Yogi Devendra
>            Assignee: Yogi Devendra
>            Priority: Minor
>
> Case 1: 
> If maxLength is set to 100 bytes. 95 bytes are already written to file. New tuple of 10 bytes arrives on input port. This tuple will be written to the same file and output will be rolled over to the next file after this tuple is written to the file. File size would be 105 bytes after the tuple is written. 
> This behaviour looks OK. No issues in this case. 
> Case 2:
> If maxLength is set to 100 bytes. 100 bytes are already written to file. 
> Expectation: This should immediately roll over to next file before next tuple is appended. 
> Current behaviour: Roll over happens only if currentLen > maxLen . i.e. it will roll over only after next tuple is appended.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)