CODEWITHSUNDEEP
pandasfor-loopscipy.stats
stvlam22
stvlam22

Reputation: 147

apply stats.percentileofscore function for multiple column

I can't get result from column ['a percent'], only column ['b percent'] get filled. can someone help ? or any reference link will be helpfull ? many thanks

test = pd.DataFrame({'a':[20,2,7,1,7,7,34],'b':[100,99,102,103,56,70,200]})
test['a percent'] = ''
test['b percent'] = ''

x = ['a','b']

for row in test.index:
    for x in x:
        test.loc[row,f'{x} percent'] = stats.percentileofscore(test[x],test.loc[row,x])/100

output:

    a   b   a percent   b percent
0   20  100 0.857143    0.571429
1   2   99              0.428571
2   7   102             0.714286
3   1   103             0.857143
4   7   56              0.142857
5   7   70              0.285714
6   34  200             1.0

Upvotes: 1

Views: 90

Answers (1)

mozway
mozway

Reputation: 261590

You overwrite your x in for x in x, use a different name:

cols = ['a','b']

for row in test.index:
    for x in cols:
        test.loc[row,f'{x} percent'] = stats.percentileofscore(test[x],test.loc[row,x])/100

That said a better, vectorial method would be:

cols = ['a', 'b']
test.update(test[cols].apply(lambda c: stats.percentileofscore(c, c))
                      .div(100).add_suffix(' percent'))

Output:

    a    b a percent b percent
0  20  100  0.857143  0.571429
1   2   99  0.285714  0.428571
2   7  102  0.571429  0.714286
3   1  103  0.142857  0.857143
4   7   56  0.571429  0.142857
5   7   70  0.571429  0.285714
6  34  200       1.0       1.0

Upvotes: 2

Related Questions