pandas str.extract multiple regex groups with OR

[perfringo]

With sample dataset one can achieve desired result as follows:

>>> df = pd.DataFrame('3" deep, 4 inches deep, 5" depth'.split(','), columns=['Depth'])
>>> df
          Depth
0         3" deep
1   4 inches deep
2        5" depth
>>> df.Depth.str.extract('(\d+)')                                     
   0
0  3
1  4
2  5

EDIT: if numeric value is needed then it probably should be int (or float) datatype. So one can do:

>>> df['Depth_number'] = df.Depth.str.extract('(\d+)').astype(int)
>>> df
            Depth  Depth_number
0         3" deep             3
1   4 inches deep             4
2        5" depth             5
>>> df.dtypes
Depth           object
Depth_number     int64
dtype: object