如何根据 pandas dataframe 中其他列中的子字符串创建新列？

Question

I have a dataframe of the following structure:

df = pd.DataFrame({
    'Substance': ['(NPK) 20/10/6', '(NPK) Guayacan 10/20/30', '46%N / O%P2O5 (Urea)', '46%N / O%P2O5 (Urea)', '(NPK) DAP Diammonphosphat; 18/46/0'],
    'value': [0.2, 0.4, 0.6, 0.8, .9]
})

    substance               value
0   (NPK) 20/10/6           0.2
1   (NPK) Guayacan 10/20/30 0.4
2   46%N / O%P2O5 (Urea)    0.6
3   46%N / O%P2O5 (Urea)    0.8
4   (NPK) DAP Diammonphosphat; 18/46/0          0.9

Now I want to create a new column with the short names of substance:

test['Short Name'] = test['Substance'].apply(lambda x: 'Urea' if 
                                         any(i in x for i in 'Urea') else '(NPK)')

There are two issues with the last line of code. First of all, the output looks like this:

    Substance               value   Short Name
0   (NPK) 20/10/6           0.2     (NPK)
1   (NPK) Guayacan 10/20/30 0.4     Urea
2   46%N / O%P2O5 (Urea)    0.6     Urea
3   46%N / O%P2O5 (Urea)    0.8     Urea
4   (NPK) DAP Diammonphosphat; 18/46/0          0.9     (NPK)

So the second entry was also labeled with Urea although it should be NPK.

Furthermore, my actual data also produces the following error, which I interestingly / annoyingly can't reproduce with the dummy data despite using the original substance names.

/var/folders/tf/hzv31v4x42q4_mnw4n8ldhsm0000gn/T/ipykernel_10743/136042259.py:5: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

Note: Since I have further substances, I will have to add more statements to the if/else loop.

Edit: The substance names need to be mapped to the following list of short names:

• Urea if Substance includes Urea
• Calcium ammonium nitrate (CAN) if Substance includes CAN
• Di-ammonium phosphate (DAP) if Substance includes DAP
• Other complex NK, NPK fertilizer for all other cases

Expected output for the sample data would be

    Substance               value   Short Name
0   (NPK) 20/10/6           0.2     (NPK)
1   (NPK) Guayacan 10/20/30 0.4     (NPK)
2   46%N / O%P2O5 (Urea)    0.6     Urea
3   46%N / O%P2O5 (Urea)    0.8     Urea
4   (NPK) DAP Diammonphosphat; 18/46/0          0.9     (NPK)

Edit2: I would then like to add a statement such that I receive the following output:

    Substance               value   Short Name
0   (NPK) 20/10/6           0.2     (NPK)
1   (NPK) Guayacan 10/20/30 0.4     (NPK)
2   46%N / O%P2O5 (Urea)    0.6     Urea
3   46%N / O%P2O5 (Urea)    0.8     Urea
4   (NPK) DAP Diammonphosphat; 18/46/0          0.9     DAP

Answer 1

Try this:

df['Short Name'] = df['Substance'].str.extract(r'\((.+?)\)')

Output:

>>> df
                 Substance  value Short Name
0            (NPK) 20/10/6    0.2        NPK
1  (NPK) Guayacan 10/20/30    0.4        NPK
2     46%N / O%P2O5 (Urea)    0.6       Urea
3     46%N / O%P2O5 (Urea)    0.8       Urea
4            (NPK) 20/10/6    0.9        NPK

Answer 2

Works for me:

df['Short Name'] = df['Substance'].apply(lambda x: 'Urea' if 'Urea' in x else '(NPK)')

>>> df
                 Substance  value   Short Name
0            (NPK) 20/10/6    0.2        (NPK)
1  (NPK) Guayacan 10/20/30    0.4        (NPK)
2     46%N / O%P2O5 (Urea)    0.6         Urea
3     46%N / O%P2O5 (Urea)    0.8         Urea
4            (NPK) 20/10/6    0.9        (NPK)

regex:

import re
short = re.compile(r"\W*(urea)\W*", re.I)
df['Short Name'] = df['Substance'].apply(lambda x: 'Urea' if len(short.findall(x.lower())) else '(NPK)')

Answer 3

Not the neatest solution but at least a solution:

test['Short Name'] = test['Substance'].apply(lambda x: 'Urea' if 'Urea' in x else 'DAP' if 'DAP' in x else '(NPK)')