pandas split string column on any character that is not an alphabet

I have dataframe something like below

a    str_col
1    ABC*EFG
2    DDC/DSD
3.   sew^sds 
...

I want to split them on non alphabet and into a list. Desired df is as follows

a    str_col.   new_col
1    ABC*EFG.   [ABC, EFG]
2    DDC/DSD.   [DDC, DSD]
3.   sew^sds    [sew, sds]
...

I’ve tried

• df['str_col'].str.split('^[a-zA-Z]+') but it created something like [, *EFG]`

>Solution :

You can use [^a-zA-Z], or \W+ (equivalent to [^a-zA-Z0-9_]) that should also work in your case:

df['new_col'] = df['str_col'].str.split(r'[^a-zA-Z]+')

df['new_col'] = df['str_col'].str.split(r'\W+')

Output:

     a  str_col     new_col
0  1.0  ABC*EFG  [ABC, EFG]
1  2.0  DDC/DSD  [DDC, DSD]
2  3.0  sew^sds  [sew, sds]

^[a-zA-Z]+ failed because ^ is an anchor to the start of the string when outside of […].