You are viewing a plain text version of this content. The canonical link for it is here.

Posted to jira@arrow.apache.org by "Jonathan Keane (Jira)" <ji...@apache.org> on 2021/09/01 20:36:00 UTC

[jira] [Updated] (ARROW-13852) [R] Handle Dataset schema metadata in ExecPlan

     [ https://issues.apache.org/jira/browse/ARROW-13852?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jonathan Keane updated ARROW-13852:
-----------------------------------
    Priority: Blocker  (was: Major)

> [R] Handle Dataset schema metadata in ExecPlan
> ----------------------------------------------
>
>                 Key: ARROW-13852
>                 URL: https://issues.apache.org/jira/browse/ARROW-13852
>             Project: Apache Arrow
>          Issue Type: New Feature
>          Components: R
>            Reporter: Neal Richardson
>            Priority: Blocker
>              Labels: query-engine
>             Fix For: 6.0.0
>
>
> See https://github.com/apache/arrow/pull/10992#issuecomment-910598194. Scanner would preserve the schema metadata but ExecPlan does not, and that makes sense in the general case. But this may result in unexpected behavior in R where we keep R attributes in the schema metadata. 
> We could probably determine in ExecPlan$Build which fields in the resulting Table are just carried forward from the underlying Dataset and pull in metadata for them to the result.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)