R - How to create a vector with propagated dates - r

Question

I am trying to create a vector (dateVec) which contains the dates in the column Date propagated by the number of days in column Days. I cannot understand why the code that I created is not working. Dates are in Date format.
> for ( i in mydata[,1] ) {
> dateVec = mydata [,1] + 0 : mydata [,2] }
The data has much more rows, here is a sample as an example:
Date (mydata[,1]) -- Days (mydata[,2])
10/05/2017 ---------- 3
05/05/2017 ---------- 2
The result that I would expect for dateVec would be:
(10/05/2017, 11/05/2017, 12/05/2017, 13/05/2017, 05/05/2017, 06/05/2017, 07/05/2017, ...)

Answer 1

There are a few issues here why your code isn't working.
For loop: Here, your i needs a series of integers to iterate through. As you
have it now, you are trying to loop from i = 1 to "10/05/2017" and "05/05/2017".
A more useful way is to use seq_along to generate a sequence of integers from
1 to the length of the object passed through to seq_along.
dateVec is not indexed, so that you are overwriting dateVec for each
iteration of your loop
Variable length of days. For the first date, you are generating a sequence 3
days long and for the second date, 2 days. You will need a data structure that can handle variable length element such as a list.
To modify your existing code:
mydata <- data.frame(Date = as.Date(c("10/05/2017", "05/05/2017"),
format = "%d/%m/%Y"), Days = c(3, 2))
dateVec <- list()
for (i in seq_along(mydata[, 1])) {
dateVec[[i]] = mydata [i, 1] + 0 : mydata [i, 2]
}
res <- do.call("c", dateVec)
A more r idiomatic approach is to pass the starting date and length of time in parallel using mapply to return a list, which is then concatenated to a vector of dates
res <- do.call("c", mapply(function(x, y) seq(from = x, length.out = y,
by = "1 day"), x = mydata[["Date"]], y = mydata[["Days"]]))

Answer 2

Here's a clunky solution:
library("lubridate")
mydata = data.frame(Date = dmy(c("10/05/2017", "05/05/2017")),
Days = c(3,2))
dateVec = dmy(character())
for(i in 1:length(mydata$Date)){
dateVec = c(dateVec,mydata$Date[i])
for(j in 1:mydata$Days[i]){
dateVec = c(dateVec, mydata$Date[i]+j)
}
}
Note that this uses the lubridate package and doesn't format the dates quite how you did. I also found it interesting that I had to initialize dateVec as a date object. Initially I tried dateVec = c() but R tried to coerce to numeric.