CODEWITHSUNDEEP
pythonpandasdataframe
Demaunt
Demaunt

Reputation: 1243

filter DataFrame using dictionary

I am new to pandas and python. I want to use dictionary to filter DataFrame

import pandas as pd
from pandas import DataFrame

df = DataFrame({'A': [1, 2, 3, 3, 3, 3], 'B': ['a', 'b', 'f', 'c', 'e', 'c'], 'D':[0,0,0,0,0,0]})
my_filter = {'A':[3], 'B':['c']}

When I call

df[df.isin(my_filter)]

I get

     A    B   D
0  NaN  NaN NaN
1  NaN  NaN NaN
2  3.0  NaN NaN
3  3.0    c NaN
4  3.0  NaN NaN
5  3.0    c NaN

What I want is

     A    B   D
3  3.0    c   0
5  3.0    c   0

I dont want to add "D" in the dictionary, I want to get rows that has proper values in A and B clumns

Upvotes: 4

Views: 5305

Answers (1)

jezrael
jezrael

Reputation: 862601

You can sum of True by columns and then compare with 2:

print (df.isin(my_filter).sum(1) == 2)
0    False
1    False
2    False
3     True
4    False
5     True
dtype: bool

print (df[df.isin(my_filter).sum(1) == 2])
   A  B  D
3  3  c  0
5  3  c  0

Another solution with first filter only columns with condition A and B with all for checking both True by columns:

print (df[df[['A','B']].isin(my_filter).all(1)])
   A  B  D
3  3  c  0
5  3  c  0

Thank you MaxU for more flexible solution:

print (df[df.isin(my_filter).sum(1) == len(my_filter.keys())])
   A  B  D
3  3  c  0
5  3  c  0

Upvotes: 6

Related Questions