[
https://issues.apache.org/jira/browse/ARROW-2298?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16646805#comment-16646805
]
Jakub Okoński commented on ARROW-2298:
--------------------------------------
Sounds great, but I must admit I'm not intimately familiar with floating point
numbers.
How could one detect problems with this conversion?
{quote}{{In [2]: np.array([2**64-1], dtype=np.uint64)}}
{{Out[2]: array([18446744073709551615], dtype=uint64)}}
{{In [3]: np.array([2**64-1], dtype=np.uint64).astype(np.float64)}}
{{Out[3]: array([1.84467441e+19])}}
{{In [4]: np.array([2**64-1],
dtype=np.uint64).astype(np.float64).astype(np.uint64)}}
{{Out[4]: array([0], dtype=uint64)}}
{quote}
Is this something you can implement quickly? If not, can you give me some
pointers so that I can implement it in pyarrow?
> [Python] Add option to not consider NaN to be null when converting to an
> integer Arrow type
> -------------------------------------------------------------------------------------------
>
> Key: ARROW-2298
> URL: https://issues.apache.org/jira/browse/ARROW-2298
> Project: Apache Arrow
> Issue Type: Improvement
> Components: Python
> Reporter: Wes McKinney
> Priority: Major
> Fix For: 0.12.0
>
>
> Follow-on work to ARROW-2135
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
