R - How to convert column of quarters to dates? - r

Question

I have a csv that contains "Period", which are quarters, and "Percent". After reading the data into R, the "Period" column is "chr" and "Percent" column is "num". I want to change the quarter values to dates, so:
for (i in 1:length(sloos_tighten$Period)) {
sloos_tighten$Period[i] <- paste("Q", substring(sloos_tighten$Period[i], 6), "/", substring(sloos_tighten$Period[i], 1, 4), sep = "")
sloos_tighten$Period[i] <- as.Date(as.yearqtr(sloos_tighten$Period[i], format = "Q%q/%Y"))
}
where the first line in the for-loop changes the format of the quarter to be readable by as.yearqtr, and the second line changes the quarter to a date. The first line works as intended, but the second line converts the date to a four-digit number. I think this is because "Period" is of type "chr", but I don't know how to change it to date. I have tried to create a new column with type date, but I cannot find any resource online that explains it. Any help is appreciated. Thanks in advance.
> dput(head(sloos_tighten, 10))
structure(list(Period = c("1990:2", "1990:3", "1990:4", "1991:1",
"1991:2", "1991:3", "1991:4", "1992:1", "1992:2", "1992:3"),
`Large and medium` = c(54.4, 46.7, 54.2, 38.6, 20, 18.6,
16.7, 10, 3.5, -3.4), Small = c(52.7, 33.9, 40.7, 31.6, 6.9,
8.8, 7, 0, -7.1, -1.7)), row.names = c(NA, 10L), class = "data.frame")
^What the data looks like after import

Answer 1

The literal for loop is fine in a sense, but unfortunately there are two problems here:
There is a class problem here: if $Period is a string, then when you reassign one of its values with something of Date class, the date is then converted into a string. This is because in R data.frame's, with few exceptions, all values in a column must be the same type. That's because a column is (almost always) a vector, and R treats vectors as homogenous.
You can get around this by pre-allocating a vector of type Date and assigning it piecemeal:
newdate <- rep(Sys.Date()[NA], nrow(sloos_tighten)) # just to get the class right
for (i in 1:length(sloos_tighten$Period)) {
tmp <- paste("Q", substring(sloos_tighten$Period[i], 6), "/", substring(sloos_tighten$Period[i], 1, 4), sep = "")
newdate[i] <- as.Date(as.yearqtr(tmp, format = "Q%q/%Y"))
}
(But please, don't use this code, look at #2 below first.)
Not a problem per se, but an efficiency: R is good at doing things as a whole vector. If you reassign all of $Period in one step, then all is faster.
sloos_tighten$Period <-
as.Date(
paste0(substring(sloos_tighten$Period, 6),
"/", substring(sloos_tighten$Period, 1, 4)),
format = "%q/%Y")
This switches from paste(.., sep="") to paste0, a convenience function. Then, it removes the leading "Q" since really we don't keep it around, so why add it (other than perhaps declarative code). Last, it does a whole vector of strings at once.
(This is taking the data sight-unseen, so untested.)