Current version does not work with numpy 2.0

[jamescooke]

Check list

• I have checked that this issue has not already been reported.
• I have confirmed this bug exists on the latest version of pandera.
• (optional) I have confirmed this bug exists on the main branch of pandera.

Current behaviour

Installing latest pandera and attempting to import will cause the error.

In a clean virtualenv with Python 3.12, install Panderas and then run the repl:

pip install pandera
python

Now just import pandera:

import pandera

And the following is received:

>>> import pandera
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/tmp/tmp.dMYyAaK7ge/venv312/lib/python3.12/site-packages/pandera/__init__.py", line 5, in <module>
    import pandera.backends
  File "/tmp/tmp.dMYyAaK7ge/venv312/lib/python3.12/site-packages/pandera/backends/__init__.py", line 4, in <module>
    import pandera.backends.base.builtin_checks
  File "/tmp/tmp.dMYyAaK7ge/venv312/lib/python3.12/site-packages/pandera/backends/base/builtin_checks.py", line 14, in <module>
    from pandera.api.checks import Check
  File "/tmp/tmp.dMYyAaK7ge/venv312/lib/python3.12/site-packages/pandera/api/checks.py", line 17, in <module>
    from pandera.strategies.base_strategies import SearchStrategy
  File "/tmp/tmp.dMYyAaK7ge/venv312/lib/python3.12/site-packages/pandera/strategies/__init__.py", line 3, in <module>
    from pandera.strategies.pandas_strategies import *
  File "/tmp/tmp.dMYyAaK7ge/venv312/lib/python3.12/site-packages/pandera/strategies/pandas_strategies.py", line 42, in <module>
    from pandera.engines import numpy_engine, pandas_engine
  File "/tmp/tmp.dMYyAaK7ge/venv312/lib/python3.12/site-packages/pandera/engines/pandas_engine.py", line 757, in <module>
    np.string_,
    ^^^^^^^^^^
  File "/tmp/tmp.dMYyAaK7ge/venv312/lib/python3.12/site-packages/numpy/__init__.py", line 397, in __getattr__
    raise AttributeError(
AttributeError: `np.string_` was removed in the NumPy 2.0 release. Use `np.bytes_` instead.. Did you mean: 'strings'?

This appears to be the source of the problem:

pandera/pandera/engines/pandas_engine.py

Line 770 in efbaa91

np.string_,

Expected behaviour

Pandera should be importable and usable.

Desktop (please complete the following information):

• OS: Linux
• Browser: n/a
• Version: python 3.12, with pandas==2.2.2, pandera==0.19.3, numpy==2.0.0

[gabrieldinizsm]

Facing the same issue using pandera[polars]

OS: Windows 10
Browser: n/a
Version: python 3.12.1, with pandas==2.2.2, pandera==0.19.3, numpy==2.0.0

[MichaelMachu]

We have the same issue, right now only option is to downgrade numpy to the previous version.

For us we have chosen the numpy==1.26.4 which works.

[jamescooke]

Oh that's a good point @MichaelMachu - I also found a work around by pinning numpy<2 in my project. Then Pandera works fine again.

[cosmicBboy]

This is a duplicate of #1656. fyi, this PR should address this issue: #1690

[jamescooke]

@cosmicBboy Sorry for the duplicate 😔 close this off as required 🙏🏻.