pandas-dev
BUG: eval fails for ExtensionArray
Pandas version checks
-
[X] I have checked that this issue has not already been reported.
-
[X] I have confirmed this bug exists on the latest version of pandas.
-
[X] I have confirmed this bug exists on the main branch of pandas.
Reproducible Example
import pandas as pd
df = pd.DataFrame({'a': pd.array([1, 2, 3]), 'b': pd.array([4, 5, 6])})
# this works as expected
print(df['a'] / df['b'])
"""
0 0.25
1 0.4
2 0.5
dtype: Float64
"""
# this is not working
print(df.eval('a / b'))
"""
TypeError: Cannot interpret 'Int64Dtype()' as a data type
"""
Issue Description
This issue is coming from pint-pandas#137. ExtensionArray is used for this project and df.eval inlcuding division (\) fails with TypeError.
I was able to able to reproduce the pint-pandas issue with pandas build in IntegerArray as shown in the above example
Traceback (most recent call last):
File "c:\CALC\DEV\pint pandas debug\test.py", line 17, in <module>
print(df.eval('a / b'))
^^^^^^^^^^^^^^^^
File "C:\Users\laurent.mutricy\AppData\Local\miniconda3\Lib\site-packages\pandas\core\frame.py", line 4738, in eval
return _eval(expr, inplace=inplace, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\laurent.mutricy\AppData\Local\miniconda3\Lib\site-packages\pandas\core\computation\eval.py", line 340, in eval
parsed_expr = Expr(expr, engine=engine, parser=parser, env=env)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\laurent.mutricy\AppData\Local\miniconda3\Lib\site-packages\pandas\core\computation\expr.py", line 809, in __init__
self.terms = self.parse()
^^^^^^^^^^^^
File "C:\Users\laurent.mutricy\AppData\Local\miniconda3\Lib\site-packages\pandas\core\computation\expr.py", line 828, in parse
return self._visitor.visit(self.expr)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\laurent.mutricy\AppData\Local\miniconda3\Lib\site-packages\pandas\core\computation\expr.py", line 413, in visit
return visitor(node, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\laurent.mutricy\AppData\Local\miniconda3\Lib\site-packages\pandas\core\computation\expr.py", line 419, in visit_Module
return self.visit(expr, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\laurent.mutricy\AppData\Local\miniconda3\Lib\site-packages\pandas\core\computation\expr.py", line 413, in visit
return visitor(node, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\laurent.mutricy\AppData\Local\miniconda3\Lib\site-packages\pandas\core\computation\expr.py", line 422, in visit_Expr
return self.visit(node.value, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\laurent.mutricy\AppData\Local\miniconda3\Lib\site-packages\pandas\core\computation\expr.py", line 413, in visit
return visitor(node, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\laurent.mutricy\AppData\Local\miniconda3\Lib\site-packages\pandas\core\computation\expr.py", line 535, in visit_BinOp
return self._maybe_evaluate_binop(op, op_class, left, right)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\laurent.mutricy\AppData\Local\miniconda3\Lib\site-packages\pandas\core\computation\expr.py", line 502, in _maybe_evaluate_binop
res = op(lhs, rhs)
^^^^^^^^^^^^
File "C:\Users\laurent.mutricy\AppData\Local\miniconda3\Lib\site-packages\pandas\core\computation\expr.py", line 538, in <lambda>
return lambda lhs, rhs: Div(lhs, rhs)
^^^^^^^^^^^^^
File "C:\Users\laurent.mutricy\AppData\Local\miniconda3\Lib\site-packages\pandas\core\computation\ops.py", line 528, in __init__
if not isnumeric(lhs.return_type) or not isnumeric(rhs.return_type):
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\laurent.mutricy\AppData\Local\miniconda3\Lib\site-packages\pandas\core\computation\ops.py", line 512, in isnumeric
return issubclass(np.dtype(dtype).type, np.number)
^^^^^^^^^^^^^^^
TypeError: Cannot interpret 'Int64Dtype()' as a data type
Looking at the traceback it seams that pandas\core\computation\ops.isnumeric() does not handle properly dtypes from ExtensionArray class (or subclass). It was suggested to consider _is_numeric class attribute.
def isnumeric(dtype) -> bool:
- return issubclass(np.dtype(dtype).type, np.number)
+ return getattr(dtype, '_is_numeric', False) or issubclass(np.dtype(dtype).type, np.number)
probably linked to #21374
Expected Behavior
df.eval('a / b') should give the same result than df['a'] / df['b']
Installed Versions
INSTALLED VERSIONS
commit : 4fb94bb09abbfcae536744740b00c987a04eed23 python : 3.11.4.final.0 python-bits : 64 OS : Windows OS-release : 10 Version : 10.0.19045 machine : AMD64 processor : Intel64 Family 6 Model 186 Stepping 2, GenuineIntel byteorder : little LC_ALL : None LANG : en_US.UTF-8 LOCALE : fr_FR.cp1252
pandas : 3.0.0.dev0+983.g4fb94bb09a numpy : 1.26.4 pytz : 2024.1 dateutil : 2.8.2 setuptools : 67.8.0 pip : 23.1.2 Cython : None pytest : 8.2.0 hypothesis : None sphinx : None blosc : None feather : None xlsxwriter : None lxml.etree : None html5lib : None pymysql : None psycopg2 : None jinja2 : None IPython : None pandas_datareader : None adbc-driver-postgresql: None adbc-driver-sqlite : None bs4 : None bottleneck : None fastparquet : None fsspec : None gcsfs : None matplotlib : 3.8.3 numba : None numexpr : 2.10.0 odfpy : None openpyxl : 3.1.2 pyarrow : 15.0.0 pyreadstat : None python-calamine : None pyxlsb : None s3fs : None scipy : 1.12.0 sqlalchemy : None tables : None tabulate : 0.9.0 xarray : 2024.2.0 xlrd : None zstandard : 0.19.0 tzdata : 2023.3 qtpy : None pyqt5 : None
Thanks for the report! Further investigations and PRs to fix are welcome.
Hey @mutricyl @rhshadrach , thank you for mentioning the issue properly and in detailed manner. I have looked into it and have issued a PR as well, do check if any issues in the PR let me know I will work on it, Thanks once again. Cheers
