[ https://issues.apache.org/jira/browse/ARROW-2124?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16359989#comment-16359989 ]
Uwe L. Korn commented on ARROW-2124: ------------------------------------ This issues seems to be fixed with 0.9.0, nevertheless I will add an explicit unit test for this. > [Python] ArrowInvalid raised if the first item of a nested list of numpy > arrays is empty > ---------------------------------------------------------------------------------------- > > Key: ARROW-2124 > URL: https://issues.apache.org/jira/browse/ARROW-2124 > Project: Apache Arrow > Issue Type: Bug > Components: Python > Affects Versions: 0.8.0 > Reporter: George Sakkis > Priority: Major > Fix For: 0.9.0 > > > See example below: > {noformat} > In [1]: import numpy as np > In [2]: import pandas as pd > In [3]: import pyarrow as pa > In [4]: num_lists = [[2,3,4], [3,6,7,8], [], [2]] > In [5]: series = pd.Series([np.array(s, dtype=float) for s in num_lists]) > In [6]: pa.array(series) > Out[6]: > <pyarrow.lib.ListArray object at 0x7f0db8ad1688> > [ > [2.0, > 3.0, > 4.0], > [3.0, > 6.0, > 7.0, > 8.0], > [], > [2.0] > ] > In [7]: num_lists.append([]) > In [8]: series = pd.Series([np.array(s, dtype=float) for s in num_lists]) > In [9]: pa.array(series) > Out[9]: > <pyarrow.lib.ListArray object at 0x7f0db8ad1e58> > [ > [2.0, > 3.0, > 4.0], > [3.0, > 6.0, > 7.0, > 8.0], > [], > [2.0], > [] > ] > In [10]: num_lists.insert(0, []) > In [11]: series = pd.Series([np.array(s, dtype=float) for s in num_lists]) > In [12]: pa.array(series) > --------------------------------------------------------------------------- > ArrowInvalid Traceback (most recent call last) > <ipython-input-99-fc3a903278e6> in <module>() > ----> 1 pa.array(series) > array.pxi in pyarrow.lib.array() > array.pxi in pyarrow.lib._ndarray_to_array() > error.pxi in pyarrow.lib.check_status() > ArrowInvalid: trying to convert NumPy type object but got float64 > {noformat} -- This message was sent by Atlassian JIRA (v7.6.3#76005)