pandas dataframe return first word in string for column

It's simple. Use as follows:

df['make'] = df['id'].str.split(' ').str[0]

Consider a regex solution with loc where it extracts everything before first space:

df.loc[df['make']=='', 'make'] = df['id'].str.extract('(.*) ', expand=False)

Alternatively, use numpy's where which allows the if/then/else conditional logic:

df['make'] = np.where(df['make']=='', 
                      df['id'].str.extract('(.*) ', expand=False), 
                      df['make'])

IDK why but with the part below

df.loc[df.make == '', 'make']

OR

df.loc[df['make'] == '', 'make']

I get the error - KeyError: 'make'

So instead I did (in case someone sees the same error):

df['make'] = df['id']
df['make'] = df.id.str.split().str.get(0)

Worked for me.

Use str.split and str.get and assign using loc only where df.make == ''

df.loc[df.make == '', 'make'] = df.id.str.split().str.get(0)

print df

               id    make
0      abarth 1.4  abarth
1        abarth 1  abarth
2  land rover 1.3   rover
3    land rover 2   rover
4    land rover 5   rover
5      mazda 4.55   mazda