Pandas Dataframe extract number out of inconsistent string

I would like to extract the number out of a column and replace the initial column by the numbers. Thanks for any help.

code:

d = {'col1': [1, 2, 3, 4], 'col2': ['3 A', '4', '7 B', '9F']}
df = pd.DataFrame(data=d)
print(df)

output:

   col1 col2
0     1  3 A
1     2    4
2     3  7 B
3     4   9F

desired output:

   col1 col2
0     1    3
1     2    4
2     3    7
3     4    9

>Solution :

Use str.extract:

df['col2'] = df['col2'].str.extract('(\d+)')
# use '^(\d+)' to limit to the leading number

or, for numeric type, combine with pandas.to_numeric:

df['col2'] = pd.to_numeric(df['col2'].str.extract('(\d+)', expand=False),
                           errors='coerce')

output:

   col1 col2
0     1    3
1     2    4
2     3    7
3     4    9