Pandas Series: split, modify first, join

I have a Pandas Series with the names of cities and districts:

London:Alpha
London
London:Beta
London:Delta
Paris

I want to add "_sub" at the end of each city (but not districts!) name so that the converted series looks like this:

London_sub:Alpha
London_sub
London_sub:Beta
London_sub:Delta
Paris_sub

As far as I understand, I need to split, change 1st part and rejoin the series, so I tried this chain:

names_df[0] = names_df[0] \
        .str.split(':') \
        .apply(lambda x: x[0] + '_sub') \
        .str.join(':')

But in this way I’ve lost a district part (Alpha, Beta, Delta) and th result looks ugly:

L:o:n:d:o:n:_:s:u:b

I’ve tried another way:

names_df[0] = names_df[0]\
    .str.split(':')\
    .apply(lambda x: '_sub:'.join(x))

But it doesn’t append ‘_sub’ to rows withoud districts 🙁

What am I doing wrong?

>Solution :

Use a single regex substitution instead:

s = pd.Series(['London:Alpha', 'London', 'London:Beta', 'London:Delta', 'Paris'])
s = s.str.replace(r'^([^:]+)', r'\1_sub', regex=True)

• ^([^:]+) – starting from the start of the string ^ captures a set of chars excepting : char

0    London_sub:Alpha
1          London_sub
2     London_sub:Beta
3    London_sub:Delta
4           Paris_sub
dtype: object