[
https://issues.apache.org/jira/browse/ARROW-1522?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16210658#comment-16210658
]
Philipp Moritz commented on ARROW-1522:
---------------------------------------
Yeah, I think we should handle this in a very similar way that we are handling
Tensors at the moment, i.e. place a reference inside of the RecordBatch and
then append a buffer at the end with the actual data. I'm busy with some other
things right now but potentially I can have a look at it during the weekend.
> [C++] Support pyarrow.Buffer as built-in type in pyarrow.serialize
> ------------------------------------------------------------------
>
> Key: ARROW-1522
> URL: https://issues.apache.org/jira/browse/ARROW-1522
> Project: Apache Arrow
> Issue Type: Improvement
> Reporter: Wes McKinney
> Fix For: 0.8.0
>
>
> This will enable objects with serialized representation as large byte strings
> to avoid memory copying on read (i.e. avoiding {{PyBytes_FromStringAndSize}})
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
