df = pd.DataFrame(np.random.randint(0,6,size=(1200000, 3)),
columns=list('ABC'))
df['sum'] = df[['A','B','C']].sum(axis=1)
df = df[df['sum']==5]
df = df.sample(n=100000)
I want to create a dataframe with three different columns with random numbers between 0 and 5 such that sum across column is 5.
You can use itertools permutations to find numbers between 0 and 5 with sum equal to 5 and assign the result to DataFrame
import itertools
df = pd.DataFrame([elem for elem in list(itertools.permutations(range(6), 3)) if sum(elem) == 5], columns = list('ABC'))
df['sum'] = df.sum(1)
A B C sum
0 0 1 4 5
1 0 2 3 5
2 0 3 2 5
3 0 4 1 5
4 1 0 4 5
5 1 4 0 5
6 2 0 3 5
7 2 3 0 5
8 3 0 2 5
9 3 2 0 5
10 4 0 1 5
11 4 1 0 5
You can check with np.random.multinomial
np.random.multinomial(5, [1/3.]*3, size=5) # here when you input size 5 , it only creat 5 lines
Out[38]:
array([[2, 2, 1],
[1, 2, 2],
[0, 3, 2],
[1, 1, 3],
[3, 1, 1]])
Finish the DataFrame
ary=np.random.multinomial(5, [1/3.]*3, size=5)
df=pd.DataFrame(ary,columns=['A','B','C'])
df['SUM']=df.sum(1)
df
Out[43]:
A B C SUM
0 1 2 2 5
1 2 2 1 5
2 1 3 1 5
3 1 1 3 5
4 1 2 2 5