What's a better way to use df.apply on multiple columns?

Question

After struggling with a csv file encoding I decided to do the encoding heresy of manually replacing some characters.

This is how the dataframe looks:

df = pd.DataFrame({'a' : 'bÉd encoded',
               'b' : ['foo', 'bar'] * 3,
               'c' : 'bÉd encoded too'})


              a    b                 c
0  bÉd encoded  foo  bÉd encoded too
1  bÉd encoded  bar  bÉd encoded too
2  bÉd encoded  foo  bÉd encoded too
3  bÉd encoded  bar  bÉd encoded too
4  bÉd encoded  foo  bÉd encoded too
5  bÉd encoded  bar  bÉd encoded too

If my only problem was column 'a' this function would be enough:

def force_good_e(row):
    col = row['a']
    if 'É' in col:
        col = col.replace('É','a') 
    return col

df['a'] = df.apply(force_good_e, axis=1)

But then I would need another function for column 'c'

I got an improvement with this:

def force_good_es(row, column):
    col = row[column]
    if 'É' in col:
        col = col.replace('É','a') 
    return col


df['a'] = df.apply(lambda x: force_good_es(x,'a'), axis=1)
df['c'] = df.apply(lambda x: force_good_es(x,'c'), axis=1)

But it got me wondering, is there a better way to do this?

ie eliminating the need to make one line of

df[n] = df.apply(lambda x: force_good_es(x,n), axis=1)

for each n column that needs to be fixed.

Answer 1

You could use str.replace

df['a'] = df['a'].str.replace('É','a')
df['c'] = df['c'].str.replace('É','a')

or like @wen mentioned in comments.

df = df.replace({'É':'a'},regex=True)

Answer 2

In case that character occurs in all columns but you want to replace it only in selected columns, and you want use apply :

df.iloc[:,[0,2]].apply(lambda x: x.str.replace('É','a'), axis=1)

Occurrence of É in first and third columns will be replaced by a .