PANDAS .where str.count < n str.replace

Question

I am trying to replace a string in one column with another string using .where like this:

dfn = df.where(df['name'].str.count(' ') < 2, df['name'].str.replace(' ', ' _ '), axis=1)

The right columns are found but the whole row is replaced with NaN...

Thank You

Answer 1

Use Series.mask :

df = pd.DataFrame({"name": ['aaa ss d', 'a s', 'a f g g']})
m = df['name'].str.count(' ') < 2
df['name'] = df['name'].mask(m, df['name'].str.replace(' ', ' _ '))
print (df)
       name
0  aaa ss d
1     a _ s
2   a f g g

Or DataFrame.loc :

m = df['name'].str.count(' ') < 2
df.loc[m, 'name'] = df.loc[m, 'name'].str.replace(' ', ' _ ')