Daniel Nugent created ARROW-12431: ------------------------------------- Summary: [Python] pa.array mask inverted when type is binary and value to be converted in numpy array Key: ARROW-12431 URL: https://issues.apache.org/jira/browse/ARROW-12431 Project: Apache Arrow Issue Type: Bug Reporter: Daniel Nugent
{code:python} Python 3.9.2 | packaged by conda-forge | (default, Feb 21 2021, 05:02:46) [GCC 9.3.0] on linux Type "help", "copyright", "credits" or "license" for more information. >>> import numpy as np >>> import pyarrow as pa >>> >>> pa.array(np.array([b'\x00']),type=pa.binary(1), mask = np.array([False])) <pyarrow.lib.FixedSizeBinaryArray object at 0x7fa080ca3640> [ null ] >>> pa.array(np.array([b'\x00']),type=pa.binary(1), mask = np.array([True])) <pyarrow.lib.FixedSizeBinaryArray object at 0x7fa080ca3700> [ 00 ] >>> pa.array([b'\x00'],type=pa.binary(1), mask = np.array([False])) <pyarrow.lib.FixedSizeBinaryArray object at 0x7fa083cc9520> [ 00 ] >>> pa.__version__ '3.0.0' >>> np.__version__ '1.20.1' {code} Happens both with FixedSizeBinary and variable sized binary (I was working with FixedSizeBinary). Does not happen for integers (presumably other types, didn't exhaustively check)? -- This message was sent by Atlassian Jira (v8.3.4#803005)