I am looking to find the no of alphabets in a mixed column in a pandas df. If the no of alphabets is more than 2, keep the two alphabets nearest to the number.

I am trying to get count using the below

df['count'] = len(re.findall('[a-zA-Z]', a['MyCOlumn']))

but get the below error:

TypeError: expected string or bytes-like object

Input

ABAB5395
DC5395
ABCD1234

Output

AB5395
DC5395
CD1234