CODEWITHSUNDEEP
pythonpandas
Trace R.
Trace R.

Reputation: 347

classic "ValueError: Truth value of a series is ambiguous"

This is a watered down version of my situation. I am working in a code environment I cannot modify. On the back end, there is a function that resembles add_sigma. I am needing to assign the function to execute across rows where the "value" column holds a particular string. The function itself forces me into a boolean expression. Any way around this?

import pandas as pd
import numpy as np


def add_sigma(percent):

    if percent >= 99.9997:
        return 6
    elif percent <= 99.9996 and percent >= 99.98:
        return 5
    elif percent <= 99.979 and percent >= 99.4:
        return 4
    elif percent <= 99.39 and percent >= 93.3:
        return 3
    elif percent <= 93.29 and percent >= 69.1:
        return 2
    elif percent <= 69.09 and percent >= 30.9:
        return 1
    elif percent <= 30.89 and percent >= 0:
        return 0


df = pd.DataFrame({'value': ['0-3 bucket', '4-7 bucket', '8+ bucket', '0-3 bucket', '0-3 bucket', '8+ bucket'],
                   'percentage': [27.68, 82.94, 32.26, 91.97, 99.82, 67.44]})

df['sigma'] = ''
df['sigma'] = np.where(df['value'] == '0-3 bucket', add_sigma(df['percentage']), df['sigma'])

expected output would be

        value  percentage sigma
0  0-3 bucket       27.68     0
1  4-7 bucket       82.94      
2   8+ bucket       32.26      
3  0-3 bucket       91.97     2
4  0-3 bucket       99.82     4
5   8+ bucket       67.44

Upvotes: 1

Views: 44

Answers (1)

Joe Ferndz
Joe Ferndz

Reputation: 8508

Like cs95 said, you can use conditions and call the add_sigma function. One way to do it is to include the condition inside the apply like this:

df['sigma'] = df.apply(lambda x: add_sigma(x.percentage) if x['value'] == '0-3 bucket' else '',axis=1)

The result will be:

        value  percentage sigma
0  0-3 bucket       27.68     0
1  4-7 bucket       82.94      
2   8+ bucket       32.26      
3  0-3 bucket       91.97     2
4  0-3 bucket       99.82     4
5   8+ bucket       67.44

Using .loc, you can do :

df.loc[df['value'] == '0-3 bucket' , 'sigma'] = df['percentage'].map(add_sigma)
df['sigma'] = df['sigma'].fillna('')

Upvotes: 1

Related Questions