TypeError: unhashable type: 'numpy.ndarray' when I use Series.str.len() in pandas query?

Question

I encountered an issue that when I use Series.str.len() in pandas query method, and actually all the functions for Series.str is not supported in some of my envs, but work in other envs, and all these envs have almost same version of pandas and numpy. (I'm sure Series.str.xxxxx could work in all my envs before)

Env1

Python 3.9.7

numpy==1.21.4

pandas==1.3.4

When I ran pd.DataFrame(columns=['core_text']).query("core_text.str.len()>1"), it raised

Traceback (most recent call last): File "", line 1, in File "/Users/huhon/miniconda3/envs/venv_dev/lib/python3.9/site-packages/pandas/core/frame.py", line 4060, in query res = self.eval(expr, **kwargs) File "/Users/huhon/miniconda3/envs/venv_dev/lib/python3.9/site-packages/pandas/core/frame.py", line 4191, in eval return _eval(expr, inplace=inplace, **kwargs) File "/Users/huhon/miniconda3/envs/venv_dev/lib/python3.9/site-packages/pandas/core/computation/eval.py", line 353, in eval ret = eng_inst.evaluate() File "/Users/huhon/miniconda3/envs/venv_dev/lib/python3.9/site-packages/pandas/core/computation/engines.py", line 80, in evaluate res = self._evaluate() File "/Users/huhon/miniconda3/envs/venv_dev/lib/python3.9/site-packages/pandas/core/computation/engines.py", line 120, in _evaluate _check_ne_builtin_clash(self.expr) File "/Users/huhon/miniconda3/envs/venv_dev/lib/python3.9/site-packages/pandas/core/computation/engines.py", line 36, in _check_ne_builtin_clash names = expr.names File "/Users/huhon/miniconda3/envs/venv_dev/lib/python3.9/site-packages/pandas/core/computation/expr.py", line 834, in names return frozenset(term.name for term in com.flatten(self.terms)) TypeError: unhashable type: 'numpy.ndarray'

Env2

Python 3.9.9

numpy==1.21.4

pandas==1.3.4

It works perfected.

Anyone can help? Thanks in advance!

Hong

Answer 1

The problem is probably with the versions of numexpr. This module evaluates all string expressions for Pandas like query or pd.eval.

The solution is to upgrade your version of numexpr (or remove and reinstall it).