[BUG] extract API returns empty string returned instead of None in a specific example

[galipremsagar]

Describe the bug
extract API seems to return an empty string in a specific case where it has to return None.

Steps/Code to reproduce bug

>>> s = cudf.Series(['a1', 'b2', 'c3'])
>>> s.str.extract(r'([ab])(\d)')
      0     1
0     a     1
1     b     2
2  None  None


>>> s.str.extract(r'([ab])?(\d)')
   0  1
0  a  1
1  b  2
2     3
>>> s.str.extract(r'([ab])?(\d)')[0][2]
''


>>> s.to_pandas().str.extract(r'([ab])?(\d)')
     0  1
0    a  1
1    b  2
2  NaN  3
>>> 

Expected behavior
The expected behavior is to return None instead of an empty string ''.

Environment overview (please complete the following information)

• Environment location: Docker
• Method of cuDF install: from source[branch-0.14]

Additional context
https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.Series.str.extract.html

[davidwendt]

Adding reference to similar issue #5157
The problem is the placement of the ? after the group (...)

>>> import pandas as pd
>>> ps = pd.Series(['a1', 'b2', 'c3'])
>>> ps.str.extract(r'([ab])(\d)')
     0    1
0    a    1
1    b    2
2  NaN  NaN
>>> ps.str.extract(r'([ab])?(\d)')
     0  1
0    a  1
1    b  2
2  NaN  3
>>> ps.str.extract(r'([ab]?)(\d)')
   0  1
0  a  1
1  b  2
2     3

The greedy quantifier ? is not supported for capturing group (...) in the cudf regex implementation. That is, I don't see how the internal code can return this information.

[galipremsagar]

Thanks, @davidwendt ! Closing this issue.