Filter a dataframe based on subset of columns

Question

The following is the dataframe

import pandas as pd
df = pd.DataFrame({'A' : [1, 1, 2, 2, 3, 4, 5],
                   'B' : [11, 11, 12, 12, 13,14,15],
                   'C' :[0.12232, 0.12232, 0.3455, 0.3455, 0.112, 0.567, 0.8901],
                   'D' :[False, True, True, True, True, True, True],
                   'E' :[True, True, False, True, True, True, True],
                   'F' :[False, True, False, True, True, True, True]})

   A   B        C      D      E      F
0  1  11  0.12232  False   True  False
1  1  11  0.12232   True   True   True
2  2  12  0.34550   True  False  False
3  2  12  0.34550   True   True   True
4  3  13  0.11200   True   True   True
5  4  14  0.56700   True   True   True
6  5  15  0.89010   True   True   True

using drop_duplicates with subset column list of dataframe does not work for me. please let me know, how to do this in a simple and fast way

If the values of columns A, B, C are duplicated. Please check if D, E, F are True, remove that row from the dataframe.

expected output dataframe:

   A   B        C      D      E      F
0  1  11  0.12232  False   True  False
2  2  12  0.34550   True  False  False
4  3  13  0.11200   True   True   True
5  4  14  0.56700   True   True   True
6  5  15  0.89010   True   True   True

Answer 1

We can use DataFrame.duplicated to check A,B and C + DataFrame.all to check D,E and F . Series.mul is used here to do the AND operation between both boolean Series to make the code cleaner:

m =(  df.duplicated(subset = ['A','B','C'],keep = False)
        .mul(df[['D','E','F']].all(axis=1)) )
df.loc[~m]

Output

   A   B        C      D      E      F
0  1  11  0.12232  False   True  False
2  2  12  0.34550   True  False  False
4  3  13  0.11200   True   True   True
5  4  14  0.56700   True   True   True
6  5  15  0.89010   True   True   True

Answer 2

Here's my way:

t = (df['D'] == True) & (df['E'] == True) & (df['F'] == True) # check where D, E, F are True
e = df.duplicated(subset=['A','B', 'C']) # Check duplicates in A, B, C
a = (e == True) & (e == True) # Check where duplicates are true in D, E, F
a = a.index[a].tolist() # get the index
df = df.drop(index=a) # drop duplicates True in D, E, F
print(df)

#    A   B        C      D      E      F
# 0  1  11  0.12232  False   True  False
# 2  2  12  0.34550   True  False  False
# 4  3  13  0.11200   True   True   True
# 5  4  14  0.56700   True   True   True
# 6  5  15  0.89010   True   True   True

Answer 3

Consider using helper columns to subset logically by:

df = df.assign(grp_count = lamba x: x.groupby(['A','B','C']).transform('count'),
               sum_bool = lamba x: x.reindex['D','E','F'].sum(axis=1))

sub_df = (df[(df['grp_count'] > 1 & df['sum_bool'] != 3) | (df['grp_count'] == 1)]
           .drop(columns=['grp_count', 'sum_bool']))