[jira] [Commented] (ARROW-12647) [Python][Dataset] Consider allowing projecting/scanning with a given schema

["Joris Van den Bossche (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/ARROW-12647?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17338967#comment-17338967 ] 

Joris Van den Bossche commented on ARROW-12647:
-----------------------------------------------

In Python, I could build up projection expressions from a schema, and specify that when creating a Scanner. But do we want such handling of a projection schema at the C++ level? (or is there already functionality available?)

Dataset has a {{ReplaceSchema}}, which is a bit similar in purpose, but at the Dataset level and using a different logic about what casting is acceptable (using {{CheckProjectable}})  compared to projecting while scanning (this might be something to clean up / make consistent)

> [Python][Dataset] Consider allowing projecting/scanning with a given schema
> ---------------------------------------------------------------------------
>
>                 Key: ARROW-12647
>                 URL: https://issues.apache.org/jira/browse/ARROW-12647
>             Project: Apache Arrow
>          Issue Type: Improvement
>          Components: Python
>            Reporter: David Li
>            Priority: Major
>              Labels: dataset, datasets
>
> As a follow-up to ARROW-12631, we could add a convenience to project with a schema, which would perform column selection and casting as appropriate.



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