如果列以字符串的一部分开头，Pandas 将替换并填充

Question

I was wondering whether there is a way to use the .replace() and .fillna() for columns headers that begin with part of a string.

For example:

data['Q5_A'].fillna(value='Not Applicable', inplace=True)

This will only apply .fillna() to column Q5_A.

Similarily:

data["Q5_S"] = data["Q5_S"].replace(",", "", regex=True)

This will only apply .replace() to column Q5_S.

What if I would like it to apply to any columns that have "Q5_" in the header?

Thanks!

Answer 1

Filter the required columns then do replace/fillna:

data.assign(**data.filter(like='Q5_').replace(",", "", regex=True))

Alternatively, if you want to modify the dataframe inplace, we can do:

cols = data.filter(like='Q5_').columns
data[cols] = data[cols].replace(",", "", regex=True)

Answer 2

You can do a for loop and perform all the processes you need as well.

for header in data.columns:
    if "Q5" in header:
        data[header].fillna(value='Not Applicable', inplace=True)
        data[header] = data[header].replace(",", "", regex=True)