ypsah opened a new issue, #46183:
URL: https://github.com/apache/arrow/issues/46183
### Describe the bug, including details regarding any error messages,
version, and platform.
Hi,
In some (rather specific) situations, the following filter expression:
`pyarrow.compute.field("x").isin([0.0])` incorrectly filters out matching rows.
Here is a minimal reproducer:
```python
from tempfile import TemporaryDirectory
import pyarrow.compute
import pyarrow.dataset
import pyarrow.parquet
data = pyarrow.table({"x": [0.0]})
f0 = pyarrow.compute.field("x") == 0.0
f1 = pyarrow.compute.field("x").isin([0.0])
with TemporaryDirectory() as tmpdir:
pyarrow.parquet.write_to_dataset(data, tmpdir)
assert data == pyarrow.dataset.dataset(tmpdir).to_table()
assert data == pyarrow.dataset.dataset(tmpdir).filter(f0).to_table()
assert data == (actual :=
pyarrow.dataset.dataset(tmpdir).filter(f1).to_table()), actual
```
Output:
```
Traceback (most recent call last):
File "/tmp/tmp.iectDMOJmW/test.py", line 18, in <module>
assert data == (actual :=
pyarrow.dataset.dataset(tmpdir).filter(f1).to_table()), actual
AssertionError: pyarrow.Table
x: double
----
x: [[]]
```
This does not happen if `x` contains other values that are not `0.0` (
`pyarrow.table({"x": [0.0, 1.0]})` --> ✅ | `pyarrow.table({"x": [0.0, 0.0,
0.0]})` --> 💥).
### Component(s)
Python, C++
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
