Closed
Description
-
I have checked that this issue has not already been reported.
-
I have confirmed this bug exists on the latest version of pandas.
-
(optional) I have confirmed this bug exists on the master branch of pandas.
Problem description
Looks like a change in numpy broke our read_stata code, so now we have a bunch of read_stata test errors popping up during CI in the numpy_dev pipeline. Maybe this Numpy PR somehow broke things? I don't think any other PR merged during the last 24 hours could have possibly done so, although I might be missing something.
Error example:
pandas/tests/io/test_stata.py:52:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
pandas/io/stata.py:1928: in read_stata
data = reader.read()
pandas/io/stata.py:1646: in read
cols_ = np.where(self.dtyplist)[0]
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
args = ([<class 'numpy.int8'>, <class 'numpy.int16'>, <class 'numpy.int32'>, <class 'numpy.float32'>, <class 'numpy.float64'>, <class 'numpy.float32'>, ...],)
kwargs = {}
relevant_args = ([<class 'numpy.int8'>, <class 'numpy.int16'>, <class 'numpy.int32'>, <class 'numpy.float32'>, <class 'numpy.float64'>, <class 'numpy.float32'>, ...], None, None)
> ???
E ValueError: invalid __array_struct__