[
https://issues.apache.org/jira/browse/ARROW-2328?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16407264#comment-16407264
]
ASF GitHub Bot commented on ARROW-2328:
---------------------------------------
Adriandorr commented on issue #1766: ARROW-2328: [C++] Writing a slice with
feather ignores the offset.
URL: https://github.com/apache/arrow/pull/1766#issuecomment-374799805
On further investigation, there is a second bug, in that the null bitmap
isn't aligned with the offset either. Hence the SliceStringsRoundTrip test
fails.
Looking at the TableWriter code, to fix this will require creating a new
null bitmap, which will require a MemoryPool and hence a change in the api. :(.
Without the ability to have nulls, writing sliced tables to feather seems to
be mostly broken. I could add the pool as an optional argument, or what would
be a preferred approach? I'll have a think tomorrow.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
> Writing a slice with feather ignores the offset
> -----------------------------------------------
>
> Key: ARROW-2328
> URL: https://issues.apache.org/jira/browse/ARROW-2328
> Project: Apache Arrow
> Issue Type: Bug
> Components: C++
> Affects Versions: 0.8.0
> Reporter: Adrian
> Priority: Major
> Labels: pull-request-available
> Fix For: 0.9.0
>
> Original Estimate: 0.5h
> Remaining Estimate: 0.5h
>
> Writing a slice from row n of length m of an array to feather would write the
> first m rows, instead of the rows starting at n.
> I've created a pull request with tests and fix here:
> [Pullrequest#1766|https://github.com/apache/arrow/pull/1766]
>
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
