Changing the year of a slice/series obtained from a pandas dataframe-CodePudding

I got a large dataset which is the imported schedule (spanning multiple years) for my team. I cleaned the data (made it long instead of wide), however I encounter a problem.

First an explanation of the data:

'year' and 'period' are obtained from the split sheetname. Both strings.
'week' the week of the year, obtained from the roster. Float.
'date' converted from string, for which I wrote a function as the dates were in Dutch and needed to be normalized, no year was defined so therefore the year from the first column is used. After processing; datetime format.
'shift' the type of shift it belongs to. S1 > early, S2 > late, S3 > night.
Each rule is assigned to one of my employees, those names are erased for privacy reasons.
I've written a class with several methods that apply rules our government enforces on schedules.

Now my problem:

As you can see: entries 1137 and 1138 should belong to the year 2022. But how do I change this easily? I tried:

for week, date in prepocessed_data_merged[['week', 'date']].values:
    # There are always more than 52 weeks in a year.
    # If the month of the date in week 52 is 1 (Jan), then something is wrong.
    if (week == 52) & (date.month == 1):
        prepocessed_data_merged.loc[(prepocessed_data_merged['week'] == week)
                                    & (prepocessed_data_merged['date']), 'date'] = ???

But as you might expect this returns a series since there are three shifts on a day, so three entries of a date that need their year changed. So, how does one change the year of a selected series/slice, simultaneously changing it in the dataframe?

I know I can use: dt.replace(year=current_year 1) but how do I enforce this replace on this selected series in the preprocessed_data DF? Thanks in advance!

CodePudding user response：

Have you tried:

cond = prepocessed_data_merged['week'].eq(52) & prepocessed_data_merged['date'].dt.month.eq(1)
prepocessed_data_merged.loc[cond, 'date']  = pd.DateOffset(years=1)

CodePudding user response：

I applied the code below, which solved my problem. If there are suggestions for making this code simpler, please do! I like to learn.

for week, date in prepocessed_data_merged[['week', 'date']].values:
    # There are always more than 52 weeks in a year.
    # If the month of the date in week 52 is 1 (Jan), something is wrong.
    if (week >= 52) & (date.month == 1):
        cond = (prepocessed_data_merged['week'] == week) & (prepocessed_data_merged['date'] == date)
        idx = prepocessed_data_merged[cond].index
        for i in idx:
            prepocessed_data_merged.loc[i, 'date'] = date.replace(year=date.year 1)

    # If the month of the date is 12 (Dec) and the week is 1, something is wrong.
    if (week == 1) & (date.month == 12):
        cond = (prepocessed_data_merged['week'] == week) & (prepocessed_data_merged['date'] == date)
        idx = prepocessed_data_merged[cond].index
        for i in idx:
            prepocessed_data_merged.loc[i, 'date'] = date.replace(year=date.year-1)