Shift dates on the fly based on a condition using pandas

Question

I have a dataframe like as shown below

df1 = pd.DataFrame({'person_id': [11, 21, 31, 41, 51],
                        'date_1': ['12/30/1961', '05/29/1967', '02/03/1957', '7/27/1959', '01/13/1971'],
                        'backward': [0,120,1,2,3],
                        'forward':  [0,12,1,0,0]})
df1['date_1'] = pd.to_datetime(df1['date_1'])

What I would like to do is

1) Shift forward date_1 column based on values from the forward column (add dates).

2) if forward column has a value of 0, then pick the value from backward and shift the date backward (subtract dates)

3) if both forward and backward are 0, it will not make any difference to dates as we add/subtract 0 days to date. For coding purposes, if you still wanna know which column to pick then we can consider the backward column.

This is what I tried

df1['new_date'] = df1['date_1'].sub(pd.to_timedelta(df1['forward'],unit='d'))
df1['bool'] = df1['date_1'].eq(df1['new_date']) # i do this to find out which rows aren't changed. So I can look at backward column to subtract days for these records
df1.loc[df1.bool == True, "new_date"] = df1['date_1'].sub(pd.to_timedelta(df1['backward'],unit='d'))

I expect my output to be like as shown below. As you can see for the last 2 rows and the 1st row, the forward column had a value of 0, so we pick values from the backward column and subtract the dates

Answer 1

you can use mask on forward to replace the 0s by minus backward column and then add to_timedelta this series to the date

df1['new_date'] = (df1['date_1'] 
                   + pd.to_timedelta(df1['forward'].mask(lambda x: x.eq(0), 
                                                         -df1['backward']), 
                                     unit='d')
                  )
print (df1)
   person_id     date_1  backward  forward   new_date
0         11 1961-12-30         0        0 1961-12-30
1         21 1967-05-29       120       12 1967-06-10
2         31 1957-02-03         1        1 1957-02-04
3         41 1959-07-27         2        0 1959-07-25
4         51 1971-01-13         3        0 1971-01-10