TST: Add any_string_dtype parameterization to test

[mroeschke]

• tests added / passed
• Ensure all linting tests pass, see here for how to run them

xref #44876 (comment)

[jbrockmendel]

should this affect expected.dtype?

[mroeschke]

Not sure cc @simonjayhawkins

We don't set the expected dtype in the rest of the tests.

[simonjayhawkins]

only if we have expand=True (the default) we get a DataFrame with the columns with the same dtype otherwise the result is an object Series of tuples of python strings.

import pandas as pd

print(pd.__version__)

s = pd.Series(["a__b__c", "c__d__e", np.nan, "f__g__h", None], dtype="string[pyarrow]")
print(s)

result1= s.str.partition(sep = "__", expand=False)
print(result1)

result2 = s.str.partition(sep="__", expand=True)
print(result2)

print(result2.dtypes)

print(repr(result2[0].dtype))

1.4.0.dev0+1658.g05b85037ab
0    a__b__c
1    c__d__e
2       <NA>
3    f__g__h
4       <NA>
dtype: string
0    (a, __, b__c)
1    (c, __, d__e)
2             <NA>
3    (f, __, g__h)
4             <NA>
dtype: object
      0     1     2
0     a    __  b__c
1     c    __  d__e
2  <NA>  <NA>  <NA>
3     f    __  g__h
4  <NA>  <NA>  <NA>
0    string
1    string
2    string
dtype: object
string[pyarrow]

[jbrockmendel]

OK thanks.

[simonjayhawkins]

Thanks @mroeschke