CODEWITHSUNDEEP
rtime-serieszooposixct
Migue
Migue

Reputation: 1441

Converting time on time series

I'm reading a csv file of precipitation time series, the problem is when the time is 00:00.

head(C1)
            fecha precip
1      09/04/2012      0
2 09/04/2012 0:05      0
3 09/04/2012 0:10      0
4 09/04/2012 0:20      0
5 09/04/2012 0:25      0
6 09/04/2012 0:30      0

I'm trying to obtain a zoo object, but if I change time to POSIXct, or POSIXlt:

C1$fecha <- as.POSIXct(C1$fecha, format="%d/%m/%Y %k:%M")

I get NA on problematic dates.

head(C1)
                fecha precip
1                <NA>      0
2 2012-04-09 00:05:00      0
3 2012-04-09 00:10:00      0
4 2012-04-09 00:20:00      0
5 2012-04-09 00:25:00      0
6 2012-04-09 00:30:00      0

Any hint on this? Thanks

Upvotes: 2

Views: 170

Answers (2)

G. Grothendieck
G. Grothendieck

Reputation: 269664

If x is your vector of dates and date/times then try this:

as.POSIXct(paste(x, "00:00"), format = "%m/%d/%Y %k:%M")

Upvotes: 0

Andrie
Andrie

Reputation: 179438

One solution is to find the strings that are short, i.e. don't have time information. Then concatenate 0:00 to each of these strings before converting to datetime objects:

Create sample data:

x <- c("09/04/2012", "09/04/2012 0:05", "09/04/2012 0:10", "09/04/2012 0:20", 
       "09/04/2012 0:25", "09/04/2012 0:30")

Use nchar() to identify elements that are only 10 characters in length. Then paste() 0:00 to these elements:

index <- sapply(x, nchar) == 10
x[index] <- paste0(x[index], " 0:00")

as.POSIXct(x, format="%d/%m/%Y %k:%M")

[1] "2012-04-09 00:00:00 BST" "2012-04-09 00:05:00 BST" "2012-04-09 00:10:00 BST"
[4] "2012-04-09 00:20:00 BST" "2012-04-09 00:25:00 BST" "2012-04-09 00:30:00 BST"

Upvotes: 3

Related Questions