Populate Consecutive Date by Groups

[atakzltn]

Hi,

I have a dataset as you can see in below:

WEEK	App ID	Total
1	123	5
1	123	6
1	123	7
1	456	3
1	456	3
1	456	2
2	123	1
2	123	5
2	123	6
2	456	2
2	456	4
2	456	3

I want to add Date column by groups (grouping by Week and App ID).

WEEK	App ID	Total	Date
1	123	5	2020-12-28
1	123	6	2020-12-29
1	123	7	2020-12-30
1	456	3	2020-12-28
1	456	3	2020-12-29
1	456	2	2020-12-30
2	123	1	2020-12-31
2	123	5	2021-01-01
2	123	6	2021-01-02
2	456	2	2020-12-31
2	456	4	2021-01-01
2	456	3	2021-01-02

In original dataset, I have 7 days for each App but it depends. So, I don't want to populate manually multiplying with 7. I want to automatize grouping by Week and App ID.

Thank you.

[gitdemont]

Hi,

start_date = as.Date.character("2020-12-27")
do.call(rbind, args = by(df, df[, c("Week", "App")], FUN = function(d) {
  d$Date = start_date + d$Week * 1:nrow(d)
  d
}))

EDIT: SORRY NOT WORKING

[gitdemont]

This should do it

df = as.data.frame(matrix(c(1,123,5,
                            1,123,6,
                            1,123,7,
                            1,456,3,
                            1,456,3,
                            1,456,2,
                            2,123,1,
                            2,123,5,
                            2,123,6,
                            2,456,2,
                            2,456,4,
                            2,456,3), ncol = 3, byrow = T))
colnames(df) = c("Week", "App", "Total")
start_date = as.Date.character("2020-12-27")

do.call(rbind, args = by(df, df[, c("App")], FUN = function(d) {
  d$Date = start_date + 1:nrow(d) 
  d
}))

But it is not taking into acount Week
Could you tell what is expected on a less repetitive dataset
Will there be exactly only a single day between each Date for the same App ?
If not what dictate this ?
e.g.:

df = as.data.frame(matrix(c(1,123,5,
                            1,123,6,
                            1,123,7,
                            1,456,3,
                            1,456,3,
                            2,123,1,
                            2,123,5,
                            2,123,6,
                            2,456,2,
                            2,456,4,
                            2,456,3,
                            3,123,4,
                            3,123,5,
                            4,123,3,
                            4,123,3,
                            4,456,3,
                            4,456,2), ncol = 3, byrow = T))

[atakzltn]

Could you please try with the code that is in below? I think it is work.

do.call(rbind, args = by(df, df[, c("Week","App")], FUN = function(d) {
  d$Date = start_date + 1:nrow(d) 
  d
}))

Edit: No, it does not because week is a dynamic column and if week increases, date also has to increase.