[
https://issues.apache.org/jira/browse/ARROW-7759?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17040088#comment-17040088
]
Antoine Pitrou commented on ARROW-7759:
---------------------------------------
I think this will require some nested parallelism support to avoid deadlocks
(if e.g. the dataset scanner spawns a task that calls the CSV reader which
itself calls TaskGroup::Finish).
> [C++][Dataset] Add CsvFileFormat for CSV support
> ------------------------------------------------
>
> Key: ARROW-7759
> URL: https://issues.apache.org/jira/browse/ARROW-7759
> Project: Apache Arrow
> Issue Type: Improvement
> Components: C++ - Dataset
> Reporter: Francois Saint-Jacques
> Priority: Major
>
> This should be a minimal implementation that binds 1-1 file and ScanTask for
> now. Streaming optimizations can be done in ARROW-3410.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
