rbind_list() / rbind_all() do not propagate timezone information for POSIXct

[ateucher]

Similar to #252 and #265:

dates1 <- data.frame(ID=c("a", "b", "c"), 
                     dates=structure(c(-247320000, -246196800, -245073600), 
                                     tzone = "GMT",
                                     class = c("POSIXct", "POSIXt")), 
                     stringsAsFactors=FALSE)

dates2 <- data.frame(ID=c("d", "e", "f"), 
                     dates=structure(c(-243864000, -242654400, -241444800), 
                                     tzone = "GMT",
                                     class = c("POSIXct", "POSIXt")), 
                     stringsAsFactors=FALSE)

sapply(list(dates1$dates, dates2$dates), attr, "tz")
# [1] "GMT" "GMT"

alldates <- rbind_list(dates1, dates2)

attr(alldates$dates, "tz")
# NULL

Cheers,
Andy

[romainfrancois]

I've handled the case where all tzone are the same. When not, I'm just considering that results are not compatible. Maybe instead we should align times against the first time zone when there are several involved.

[hadley]

That's already better than R which IIRC simply drops all time zones except the first.