How to use R apply without changing a function

Question

I have a basic character date to POSIXct function

charDateToPosixct <- function(chrDate, format, timeZone) { as.POSIXct(chrDate, format=format, tz=timeZone) }

I have a data frame with date as character and timezone.

chrDate <- c("4/25/2012","4/24/2012","4/16/2012","6/30/2012") timeZone <- c("US/Eastern","US/Central","US/Pacific","US/Eastern") df <- data.frame(date=chrDate,timezone=timeZone) str(df) 'data.frame': 4 obs. of 2 variables: $ date : Factor w/ 4 levels "4/16/2012","4/24/2012",..: 3 2 1 4 $ timezone: Factor w/ 3 levels "US/Central","US/Eastern",..: 2 1 3 2

I want to change the data type of date to POSIXct and so want to apply this function charDateToPosixct to each row with the format "%m/%d/%y".

GSee · Accepted Answer · 2012-07-18 23:04:55Z

Use an anonymous function in apply like this

apply(df, 1, function(x) charDateToPosixct(x[1], "%m/%d/%y", x[2])) #[1] 1587787200 1587704400 1587020400 1593489600

edit

You could convert the numeric vector you got using apply to POSIXct

as.POSIXct(apply(df, 1, function(x) charDateToPosixct(x[1], "%m/%d/%y", x[2])), origin='1970-01-01') [1] "2020-04-25 05:00:00 CDT" "2020-04-24 06:00:00 CDT" "2020-04-16 08:00:00 CDT" "2020-06-30 05:00:00 CDT"

Or, you can use lapply to get a list of POSIXct objects. Then you can combine with do.call c, or Reduce (You have to use as.character because you implicitly used stringsAsFactors=TRUE when you created the data.frame)

do.call(c, lapply(seq_len(NROW(df)), function(i) { charDateToPosixct(as.character(df$date[i]), "%m/%d/%Y", as.character(df$timezone[i])) }))

Or, if you have 2 vectors, you don't even need the data.frame

do.call(c, lapply(seq_along(chrDate), function(i) { charDateToPosixct(chrDate[i], "%m/%d/%Y", timeZone[i]) }))

Or, using Reduce

Reduce(c, lapply(seq_along(chrDate), function(i) { charDateToPosixct(chrDate[i], "%m/%d/%Y", timeZone[i]) }))

My timezone is CDT, so any of the above give me

[1] "2012-04-24 23:00:00 CDT" "2012-04-24 00:00:00 CDT" "2012-04-16 02:00:00 CDT" "2012-06-29 23:00:00 CDT"

if i use df$date1 <- apply(...) it doesn't change the data type to posixct. str(df$date1) is still numeric and not posixct.
That's the way apply works. My results match the results you get using the workaround that you said works.
The 2nd part of the question where i said Alternatively or workaround was wrong and so i omitted. sorry 4 the confusion.
From ?apply in the Value section (i.e. the section that tells you what the function returns): "the result is coerced by ‘as.vector’ to one of the basic vector types before the dimensions are set". So, that's why you get a vector of class numeric instead of POSIXct
I don't think so since you have different timezones. There are other ways to hide the loop, e.g. as.POSIXct(Vectorize("charDateToPosixct")(chrDate, "%m/%d/%y", timeZone), origin='1970-01-01')

Collectives™ on Stack Overflow

How to use R apply without changing a function

1 Answer 1

6 Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

6 Comments

Related