[jira] [Commented] (ARROW-13103) [Python] Allow to specify a dictionary type for dictionary_encode() function

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

    [ https://issues.apache.org/jira/browse/ARROW-13103?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17364906#comment-17364906 ] 

Joris Van den Bossche commented on ARROW-13103:
-----------------------------------------------

That would indeed be useful. For reference, another possible extension to {{dictionary_encode}} would be to specify the dictionary values up front (ARROW-1574)

> [Python] Allow to specify a dictionary type for dictionary_encode() function
> ----------------------------------------------------------------------------
>
>                 Key: ARROW-13103
>                 URL: https://issues.apache.org/jira/browse/ARROW-13103
>             Project: Apache Arrow
>          Issue Type: Improvement
>            Reporter: Andrei Nesterov
>            Priority: Major
>
> Functions {{[pyarrow.compute.dictionary_encode()|https://arrow.apache.org/docs/python/generated/pyarrow.compute.dictionary_encode.html]}} and {{[pyarrow.Array.dictionary_encode()|https://arrow.apache.org/docs/python/generated/pyarrow.DictionaryArray.html#pyarrow.DictionaryArray.dictionary_encode]}} don't allow to specify a desirable dictionary type.
> A dictionary type consists of value type, index type, and ordering.
>  Index type is set to `int32()` and ordering is set to `False`.
> Changing these properties requires redundant casting ([not implemented|https://issues.apache.org/jira/browse/ARROW-13102]).



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