How in Python, Pandas to do cumsum() on a column up to a specific sum value?

Question

Python

Hello, I have a question. There is a dataframe with numbers. You need to apply the cumsum() function to its column so that it gives the cumulative sum of this column, but the values ​​\u200b\u200bare added under a certain condition, after which the summation would start again. I'm trying to figure out how to do it without loops, so that it's as fast as possible, but maybe someone already knows the solution. Thanks in advance.

Note.
Condition: if the sum <= 6, the summation is in progress, otherwise the summation starts again
Input column ; Column on output
1 ; 1
3 ; 4
2 ; 65
; 5
3 ; 81
; 1
2 ; 3

Answer 1

The fastest thing that happened

vals = pd.Series([1,3,2,5,3,1,2])

fv = 6

def cumscums(vals, fv = fv):
    vals_ = []
    vals_.append(vals[0])  
    vals = vals[1:]
    for ind,line in enumerate(vals, start=1):
        if vals_[ind-1] < fv:
            vals_.append(vals_[ind-1] + line)
        else:
            vals_.append(line)
    return vals_

ccc = cumscums(vals, fv = fv)
pd.Series(ccc)

Answer 2

import pandas as pd

l = 

df = pd.DataFrame(l)
df['cumsum'] = df[1].cumsum()
df['group1'] = df[1].cumsum() // 6 # целочисленное деление как признак группы
df['group2'] = df[1] < 6

df['cum_6'] = df.groupby(['group1','group2'])[1].cumsum()

print(df)