How to change datetime timezone in R

Here is a simple way how to change datetime timezone in the R data frame. If the timezone is not present or is different, there can cause a problem in your calculations. Usually like incorrect results of time difference.

It might usually happen by combining multiple data frames. Here is one of the examples – R time-shift problem after merging data.

What is the problem?

Here is my data frame.

ndf <- structure(
  list(
    dt1 = structure(
      c(
        1601894367.723,
        1601894396.223,
        1601894367.723,
        1601894396.223
      ),
      class = c("POSIXct", "POSIXt"),
      tzone = "UTC"
    ),
    dt2 = structure(
      c(
        1601894451.737,
        1601894451.737,
        1601895005.243,
        1601895005.243
      ),
      class = c("POSIXct", "POSIXt"),
      tzone = ""
    )
  ),
  row.names = c(NA, -4L),
  class = "data.frame"
)

ndf
#                  dt1                 dt2
#1 2020-10-05 10:39:27 2020-10-05 13:40:51
#2 2020-10-05 10:39:56 2020-10-05 13:40:51
#3 2020-10-05 10:39:27 2020-10-05 13:50:05
#4 2020-10-05 10:39:56 2020-10-05 13:50:05

When I do time difference calculations, the results are incorrect. The difference in hours must be greater.

ndf$diff <- difftime(ndf$dt1, ndf$dt2, units = 'hours')

ndf
#                  dt1                 dt2              diff
#1 2020-10-05 10:39:27 2020-10-05 13:40:51 -0.02333722 hours
#2 2020-10-05 10:39:56 2020-10-05 13:40:51 -0.01542056 hours
#3 2020-10-05 10:39:27 2020-10-05 13:50:05 -0.17708889 hours
#4 2020-10-05 10:39:56 2020-10-05 13:50:05 -0.16917222 hours

I can see that every datettime column has a different timezone that causes the problem. One has a UTC timezone, the other having none.

attr(ndf$dt1, "tzone")
#[1] "UTC"

attr(ndf$dt2, "tzone")
#[1] ""

Change datetime timezone in R

You can do a timezone change like this.

ndf$dt2 <- as.POSIXct(format(ndf$dt2), tz = "UTC")

If I change the necessary datetime timezone for the column, the calculation works fine.

ndf$diff <- difftime(ndf$dt1, ndf$dt2, units = 'hours')

ndf
#                  dt1                 dt2              diff
#1 2020-10-05 10:39:27 2020-10-05 13:40:51 -0.02333722 hours
#2 2020-10-05 10:39:56 2020-10-05 13:40:51 -0.01542056 hours
#3 2020-10-05 10:39:27 2020-10-05 13:50:05 -0.17708889 hours
#4 2020-10-05 10:39:56 2020-10-05 13:50:05 -0.16917222 hours

If you want to remove the timezone in POSIXct object then leave the tz parameter empty.

ndf$dt2 <- as.POSIXct(format(ndf$dt2), tz = "")