[
https://issues.apache.org/jira/browse/TAJO-1384?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14355550#comment-14355550
]
ASF GitHub Bot commented on TAJO-1384:
--------------------------------------
Github user hyunsik commented on the pull request:
https://github.com/apache/tajo/pull/406#issuecomment-78130692
The change looks straightforward. The second patch that eliminates
unnecessary path list and unnecessary sort looks much better than the first one.
If all unit tests are passed, I'll finish the review.
> Duplicated output file path problem
> -----------------------------------
>
> Key: TAJO-1384
> URL: https://issues.apache.org/jira/browse/TAJO-1384
> Project: Tajo
> Issue Type: Bug
> Components: storage
> Reporter: Jihoon Son
> Assignee: Jihoon Son
> Fix For: 0.11
>
> Attachments: TAJO-1384.patch
>
>
> *Introduction*
> When consecutively inserting data into an existing directory, each files must
> have different names. To avoid the duplicated name problem, a unique sequence
> number is appended to the end of the file name. In addition, the max sequence
> number of pre-existing files is first found before starting writing to the
> final output directory.
> *Problem*
> When finding the max sequence number of pre-existing files, those files are
> not sorted by their sequence numbers. This finally causes the duplicated file
> name problem.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
