Dates, timezone and POSIXct

Question

I just fread a large datafile, and the "DATE" column is stored as character like O5JAN2004, 06JAN2004.The time in this datafile is matched to NewYork time, and I live in Los Angeles.

Then I use as.Date to convert character to date.

t <- as.Date(key$DATE[1], format = "%d%b%Y")
[1] "2004-01-05"

But when I use as.POSIXct(t), it returns me:

 > as.POSIXct(t)
[1] "2004-01-04 16:00:00 PST"
 > as.POSIXlt(t)
[1] "2004-01-05 UTC"

I tried several methods mentioned on website, but the result didn't change:

t <- as.Date(key$DATE[1], format = "%d%b%Y", 'PST')
t <- as.Date(key$DATE[1], format = "%d%b%Y", 'EST')
t <- as.Date(key$DATE[1], format = "%d%b%Y", tz="America/New_York")
t <- as.Date(keyi$DATE[1], format = "%d%b%Y", tz="America/Los_Angeles")
as.POSIXct(t, tz = "America/Los_Angeles")
as.POSIXct(t, tz = "America/New_York")

I want to know: what could I do so when I use as.POSIXct(t), it would return me "2004-01-05 PST" or any other timezone.

I am thinking because the Date is originally stored as character, so it wouldn't remember its original timezone, right?

I do get

as.Date(as.POSIXct(t))
> "2004-01-05"

But why would as.POSIXct(t) return the previous result? Because I also have other data files and I would get "2004-01-05 PST" using as.POSIXct(t).

Thank you!

Answer 1

Your as.POSIXlt(t) shows that as.Date is using GMT by default. But as.POSIXct uses local time by default, so there is an unwanted conversion. But you can fix this.

strptime has a tz argument to specify the base time zone. This worked for me:

t = strptime(key$DATE[1], format = "%d%b%Y", tz= "America/Los_Angeles")
as.POSIXct(t)
[1] "2004-01-05 PST"

Warning: What does and does not work as a value for tz seems to be rather odd.