Filtering pandas DF by dict-CodePudding

I have a pandas df:

and want to iterate though a dictionary of filters to isolate the desired row.

For example, filtering on the following should return row:

df =  pd.DataFrame([[10,'A','40%'],
                    [20,'B','20%'],
                    [30,'C','10%'],
                    [40,'D','30%']],
                   columns= ['Col1','Col2','%']) 
dic = {'Col1': 10, 'Col2': 'A'}

My current approach is hardcoded to accept just two filter conditions (Col1, Col2) however I need a more generalized approach that can accept N conditions (or a dict of any size).

CodePudding user response：

You can use a Series for automatic alignment and any or all depending on whether you want any or all conditions True:

dic = {'Col1': 10, 'Col2': 'A'}

# we want both Col1 AND Col2 to match
out = df[df[list(dic)].eq(pd.Series(dic)).all(1)]

# we want either Col1 OR Col2 to match (or both)
out = df[df[list(dic)].eq(pd.Series(dic)).any(1)]

output:

   Col1 Col2    %
0    10    A  40%

CodePudding user response：

I can't load your pictures but what about this approach?

conditions_dict = {'Col1': 10, 'Col2': 'A'}

def filter_function(row):
    for col, val in conditions_dict.items():
        if row[col] != val:
            return False
    return True

df[df.apply(filter_function, axis=1)]

CodePudding user response：

You can try create DataFrame from dict then merge

out = pd.DataFrame([dic]).merge(df)

print(out)

   Col1 Col2    %
0    10    A  40%