I have a data frame which has column A consist of positive and negative number. I would like to make a column B which is the sum up to the current row of column A and start from the point value in A change sign just like below. And I have no clue where to start. Any help please?

    A       B
    4       4
    4       8
    17      25
    -1      -1
    -8      -9
    -9      -18
    -45     -63
    4       4
    5       9
    9       18
    4       22

CodePudding user response：

Use groupby and cumsum

df['B'] = df['A'].groupby(df['A'].lt(0).diff().cumsum().fillna(0)).cumsum()