CODEWITHSUNDEEP
pythonpython-3.xpandasnumpy
foy
foy

Reputation: 427

Generate a Column with random Percentage that sum to one

I would like to create a random percentage column where it should sum up to 1

Current Table :

Col1    Value
A        100
B        100

Wanted Example 1 : if there are two rows, then the random value can be any number,but the total has to be 1

Col1    Value  Random  New Value
A        100    0.7    70
B        100    0.3    30
                 1

Wanted Example 2 : three rows samples

Col1    Value  Random  New Value
A        100    0.2     20
B        100    0.1     10
C        100    0.7     70
                 1

Upvotes: 3

Views: 1155

Answers (2)

Mykola Zotko
Mykola Zotko

Reputation: 17882

You can use the numpy function random.randint:

df = pd.DataFrame({'Value': [100, 100, 100]})

nums = np.random.randint(10, size=len(df))
df['Random'] = nums / nums.sum()
df['New'] = df['Value'] * df['Random']

df.loc['Sum', :] = df.sum()

Output:

     Value  Random     New
0    100.0  0.1250   12.50
1    100.0  0.3125   31.25
2    100.0  0.5625   56.25
Sum  300.0  1.0000  100.00

Upvotes: 2

Sayandip Dutta
Sayandip Dutta

Reputation: 15872

You can use np.random.dirichlet and np.around:

import pandas as pd
import numpy as np
df = pd.DataFrame({'Col1': list("ABC"), 'Value': [100]*3})
df['random'] = np.around(np.random.dirichlet
                        (np.ones(df.shape[0]),size=1)[0],
                         decimals = 1)
df['New value'] = (df['Value']*df['random']).astype(int)
print(df)

Output:

  Col1  Value  random  New value
0    A    100     0.4         40
1    B    100     0.3         30
2    C    100     0.3         30

Upvotes: 2

Related Questions