Add timedelta to date column based on conditions-CodePudding

Here is my sample data from df.to_dict()

{'Date': {0: Timestamp('2022-01-01 00:00:00'),
  1: Timestamp('2022-01-02 00:00:00'),
  2: Timestamp('2022-01-03 00:00:00'),
  3: Timestamp('2022-01-04 00:00:00')},
 'tf': {0: 'D', 1: 'W', 2: 'M', 3: 'Y'},
 'tf_int': {0: 1, 1: 3, 2: 2, 3: 2}}

//Sorry this isn't a pretty table, I couldn't get the normal df to be pasted here

I want to make a new column based that will add the required timedelta to the Date column.

For eg, for my first row, I want to add 1 day to the date (2022-1-1) so the result would be 2022-1-2

Is there any vectorised way to achieve this ? I don't using multiple .loc[] calls to accomplish this separately for days,weeks, months etc; But I would prefer the solution to be vectorised

My unsuccessful attempt with .loc[]

df.loc[df['tf'] =='D','publish_date'] = df['Date'].dt.date   dt.timedelta(days = df['tf_int'])

But this results in error because df['tf_int'] is a series and not an exact value

Edit: For my use case, adding one month will add affect only the month component, and not do anything to the date

CodePudding user response：

If need exact year and months is necessary use list comprhension solution with offsets.DateOffset:

d = {'D':'days','W':'weeks', 'M':'months', 'Y':'years'}

df['publish_date'] = [z   pd.offsets.DateOffset(**{d[y]: x}) 
                     for x, y, z in zip(df['tf_int'], df['tf'], df['Date'].dt.normalize())]
print (df)
        Date tf  tf_int publish_date
0 2022-01-01  D       1   2022-01-02
1 2022-01-02  W       3   2022-01-23
2 2022-01-03  M       2   2022-03-03
3 2022-01-04  Y       2   2024-01-04

CodePudding user response：

If you accept to have approximations for Month/Year, a vectorial solution could be something like:

mapper = {'D': 1, 'W': 7, 'M': 30, 'Y': 365}
df['Date2'] = (pd.to_datetime(df['Date'])
                 .add(pd.to_timedelta(df['tf'].map(mapper).mul(df['tf_int']),
                                      unit='D'))
              )

output:

        Date tf  tf_int      Date2
0 2022-01-01  D       1 2022-01-02
1 2022-01-02  W       3 2022-01-23
2 2022-01-03  M       2 2022-03-04
3 2022-01-04  Y       2 2024-01-04