CODEWITHSUNDEEP
pythonpandasconditional-statementspercentile
michael0196
michael0196

Reputation: 1637

Assign multiple conditional values to percentiles into a new pandas column

I have the following dataframe with percentiles and I would like to assign values between 0-9 to their respective percentiles.

This is the input:

         a
 1  0.3038
 2  0.0572
 3  0.1153
 4  0.2168
 5  0.5688
 6  0.6291
 7  0.9416
 8  0.7438
 9  0.2242
10  0.5662

And this is the desired output:

         a  b
 1  0.3038  3
 2  0.0572  0
 3  0.1153  1
 4  0.2168  2
 5  0.5688  5
 6  0.6291  6
 7  0.9416  9
 8  0.7438  7
 9  0.2242  2
10  0.5662  5

Also, if the value in column 'a' is 1.00 I would like to return the value 9. For example:

        a   b
1  1.0000   9

Thank you for your help.

Upvotes: 1

Views: 121

Answers (1)

sjw
sjw

Reputation: 6543

Updated solution

Thanks to @Peter Leimbigler for the following suggestion in the comments, which makes sure that the requirement that 1.0 maps to 9 is correctly handled:

np.floor(df['a']*10).replace({10:9})

Alternatively, sticking with the clip_upper() idea which was present in the previous solution:

np.floor(df.clip_upper(0.9) * 10).astype(int)

0.9 in the above solution can actually be any number [0.9, 1), and will work as intended. The problem in the previous solution is that 1 was used here instead, which meant that a value of exactly 1.0 was not rounded down.

I will leave the original solution below since it was accepted by OP, but as pointed out by @Peter Leimbigler, it did not correctly handle the 1.0 -> 9 special case.

Previous solution

If I understand correctly:

df['b'] = np.floor(df.clip_upper(1) * 10).astype(int)

Gives the same result as in the question and accounts for your caveat regarding a number that is exactly 1.

         a  b
1   0.3038  3
2   0.0572  0
3   0.1153  1
4   0.2168  2
5   0.5688  5
6   0.6291  6
7   0.9416  9
8   0.7438  7
9   0.2242  2
10  0.5662  5

Upvotes: 3

Related Questions