Replace string with np.nan if condition is met

August 22, 2022

I am trying to replace a string occurrence in a column if a condition is met.

My sample input dataset:

Series Name     Type  
Food            ACG
Drinks          FEG
Food at Home    BON

I want to replace the strings on the Series Name column if the strings on the Type column are either ACG or BON with nan or blank. For that I tried the following code where I used conditions with not much success.

Code:

df.loc[((df['Type'] == 'ACG') | (df['Type'] == 'BON')),
                df['Series Name'].replace(np.nan)]

Desired output:

Series Name     Type  
                ACG
                FEG
Food at Home    BON

>Solution :

Since you want to set the whole cell to nan, just do this:

df.loc[((df['Type'] == 'ACG') | (df['Type'] == 'BON')), 'Series Name'] = np.nan

Output:

  Series Name Type
0         NaN  ACG
1      Drinks  FEG
2         NaN  BON

Update:
Regarding to your question in the comments, if you only wanted to change parts of the string, you could use replace like this:

#new input
df = pd.DataFrame({
    'Series Name': ['Food to go', 'Fast Food', 'Food at Home'], 
    'Type': ['ACG', 'FEG', 'BON']
})

    Series Name Type
0    Food to go  ACG
1     Fast Food  FEG
2  Food at Home  BON

mask = df['Type'].isin(['ACG', 'BON'])
df.loc[mask, 'Series Name'] = (df.loc[mask, 'Series Name']
                               .replace(to_replace="Food", value='NEWVAL', regex=True))
print(df)

      Series Name Type
0    NEWVAL to go  ACG
1       Fast Food  FEG
2  NEWVAL at Home  BON