Trouble writing function to remove prefixes in Python

Question

I'm trying to write a function that will remove a prefix from every element of a column in a pandas dataframe. I've made a few attempts but none have seemed to work:

prefixes = ['mm10---', 'GRCh38-']
def clean_genes(column):
    for gene in CTRL_data[f'{column}']:
        for prefix in prefixes:
            if row[f"{column}"].str.startswith(f"{prefix}"):
                gene = str.replace(f"{prefix}", '', gene)
    return column

def clean_genes(column):
    for gene in CTRL_data[f"{column}"]:
        gene = gene[7:]
    return column

clean_genes(gene)

Could someone point out where these attempts have gone wrong, or how I could better write this function? The error in both cases is:

NameError                                 Traceback (most recent call last)
/var/folders/pg/d3z5dn_x0f51tlwtj7391tjh0000gn/T/ipykernel_10029/2341573264.py in <module>
     16     return column
     17 
---> 18 clean_genes(gene)

NameError: name 'gene' is not defined

EDIT: I've also looked at some other questions on this site and others, including this one which I thought was helpful (Remove specific characters from a string in Python).

`gene` isn't defined outside of the `clean_genes` function. Also, you have 2 functions named `clean_genes`. — Adid, Jan 03 '22 at 06:16
@Julien 'gene' is the name of the column in my dataframe. Wouldn't this be more or less equivalent to writing ```column=gene```? — Aaron, Jan 03 '22 at 06:18
@Adid the two functions were separate attempts - I hashed out one when running the other — Aaron, Jan 03 '22 at 06:19
No that's not how it works. Variable name `gene` and string `'gene'` are not the same thing. Computers don't understand 'more or less'... — Julien, Jan 03 '22 at 06:21

score 2 · Answer 1 · answered Jan 03 '22 at 06:22

If your question actually is "how to remove a number of prefixes from a Pandas dataframe series", then I'd maybe say

create a regular expression to match those prefixes
use .str.replace on those series

This will likely be a lot faster than a manual loop too.

import re
prefixes = ['mm10---', 'GRCh38-']

# Build a regexp that matches either of the given prefixes, anchored
# to the start of the string.
prefix_re = re.compile("^(" + "|".join(re.escape(prefix) for prefix in prefixes) + ")")

df["my_series"] = df["my_series"].str.replace(prefix_re, "")

score 1 · Answer 2 · answered Jan 03 '22 at 06:52

1

You can remove the prefix by building regular expression that matches either of your prefixes and then use the regex to replace them with empty string like this:

re = r'^(mm10\-\-\-|GRCh38\-)'
df["my_series"] = df["my_series"].str.replace(re, "")

answered Jan 03 '22 at 06:52

Babatunde Mustapha

2,131
20
21

`-` is not a special character in regexps outside a character class and doesn't need to be escaped. – AKX Jan 03 '22 at 06:57

Trouble writing function to remove prefixes in Python

2 Answers2