[
https://issues.apache.org/jira/browse/ARROW-1763?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16235716#comment-16235716
]
Wes McKinney commented on ARROW-1763:
-------------------------------------
We should probably just hash the pickled representation
> DataType should be hashable
> ---------------------------
>
> Key: ARROW-1763
> URL: https://issues.apache.org/jira/browse/ARROW-1763
> Project: Apache Arrow
> Issue Type: Improvement
> Reporter: Jeff Reback
> Priority: Normal
> Fix For: 0.8.0
>
>
> We can then use the DataType objects as keys in dictionary for example. xref
> https://github.com/ibis-project/ibis/pull/1194#discussion_r148493472
> {code}
> In [1]: import pyarrow as pa
> In [2]: pa.__version__
> Out[2]: '0.7.1'
> In [3]: pa.int8()
> Out[3]: DataType(int8)
> In [4]: hash(pa.int8())
> ---------------------------------------------------------------------------
> TypeError Traceback (most recent call last)
> <ipython-input-4-bca0e6e2f6af> in <module>()
> ----> 1 hash(pa.int8())
> TypeError: unhashable type: 'pyarrow.lib.DataType'
> {code}
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
