Python - New Column based on condition of other Column

Question

I have a dataframe column containing the following:

Audience
searchretargeting
data-capture-320x50
purchase-behavior-320x500
data-capture-728x90

I want to create a new column (Audience2) by splitting out the 'Audience' column based on the '-' delimiter, ideally would like to only keep the 1st element of the split ('data' not the 'capture-320x50')

If there is no '-' present I would like the new column to be populated with what was in 'Audience'(e.g. searchretargeting):

Audience               Audience2
siteretargeting        siteretargeting
data-capture-320x50    data

I know how to str split the Audience column, but looking to add in some type of logic to circumvent the new column being NaN when there is no '-' present in the column

df['Audience2']=df['Audience'].str.split('-').str[1]

This splits the Audience column and only retains the first element but I've been struggling with various if-else and apply-lambda statements to figure out how to pull in data that doesn't have '-' without it being NaN

score 3 · Accepted Answer · answered Jul 12 '20 at 01:18

3

Try this:

df['Audience'].str.split('-').str[0].fillna(df['Audience'])

Output:

0    searchretargeting
1                 data
2             purchase
3                 data
Name: Audience, dtype: object

answered Jul 12 '20 at 01:18

Scott Boston

147,308
15
139
187

1

This is smart ~ – BENY Jul 12 '20 at 01:33
Nice answer. Though If we are getting the first item `str[0]` than is `fillna` neccessary? I seem to be getting the original value if we use `str[0]` without the usage of `fillna`. – MasayoMusic Jul 12 '20 at 01:49
@MasayoMusic The NaN comes from the fact that '-' doesn't appear in the original string, therefore the `split` method returns a NaN, there is no list so getting str[0] is not possible. – Scott Boston Jul 12 '20 at 01:50

score 2 · Answer 2 · answered Jul 12 '20 at 01:32

Let us do fix with ffill

df.Audience.str.split('-',expand=True).ffill(axis=1).iloc[:,1]
0    searchretargeting
1              capture
2             behavior
3              capture
Name: 1, dtype: object

More info

df.Audience.str.split('-',expand=True).ffill(axis=1)
                   0                  1                  2
0  searchretargeting  searchretargeting  searchretargeting
1               data            capture             320x50
2           purchase           behavior            320x500
3               data            capture             728x90

unltd_J · Answer 3 · 2020-09-07T01:19:39.033

1

df['audience2'] = [i.split('-')[0] for i in df.audience if '-' in i else i]

This should do it for you.

edited Sep 07 '20 at 01:19

answered Jul 12 '20 at 01:19

unltd_J

476
7
14

score 1 · Answer 4 · answered Jul 12 '20 at 01:24

You could try with np.where:

df['Audience2']=np.where(df.Audience.str.contains('-'), df.Audience.str.split('-').str[0],df.Audience)

Output:

df
                    Audience          Audience2
0          searchretargeting  searchretargeting
1        data-capture-320x50               data
2  purchase-behavior-320x500           purchase
3        data-capture-728x90               data

score 1 · Answer 5 · answered Jul 12 '20 at 01:29

1

You can instead use something like this -

df['Audience2']=df['Audience'].str.split('-').str[1]
for i in range(len(df)):
  if pd.isna(df['Audience2'][i]):
    df['Audience2'][i] = df['Audience'][i]

answered Jul 12 '20 at 01:29

Rishit Dagli

1,000
8
20

Python - New Column based on condition of other Column

5 Answers5