0

Suppose I have data that looks like this:

        DATE  TIME   Col1   Col2
1 1993-01-04 34538 10.250 10.000
2 1994-01-05 34541 10.250 10.111
3 1997-03-16 34546 10.250 10.222
4 2017-11-10 34561 10.251 10.333
5 2001-08-28 34565 10.251 10.444
6 2006-04-20 34807 10.251 10.555

The 'TIME' column is formatted as seconds from midnight. How would I combine the 'DATE' and 'TIME' columns to get an xts object that looks something like this:

                       Col1   Col2
X1993.01.04.09.35.38 10.250 10.000
X1994.01.05.09.35.41 10.250 10.111
X1997.03.16.09.35.46 10.250 10.222
X2017.11.10.09.36.01 10.251 10.333
X2001.08.28.09.36.05 10.251 10.444
X2006.04.20.09.40.07 10.251 10.555
D.J. P.
  • 98
  • 8

2 Answers2

1

We can convert the 'DATE', 'TIME' column to Datetime class and convert the dataset to xts by specifying the order.by

library(xts)     
library(lubridate)
xts(df1[-(1:2)], order.by = as.POSIXct(paste(df1$DATE, 
             hms::hms(seconds_to_period(df1$TIME)))))
#                      Col1   Col2
#1993-01-04 09:35:38 10.250 10.000
#1994-01-05 09:35:41 10.250 10.111
#1997-03-16 09:35:46 10.250 10.222
#2001-08-28 09:36:05 10.251 10.444
#2006-04-20 09:40:07 10.251 10.555
#2017-11-10 09:36:01 10.251 10.333

NOTE: The index of xts needs a Datetime class object and not a formatted character class vector

data

df1 <- structure(list(DATE = c("1993-01-04", "1994-01-05", "1997-03-16", 
"2017-11-10", "2001-08-28", "2006-04-20"), TIME = c(34538L, 34541L, 
34546L, 34561L, 34565L, 34807L), Col1 = c(10.25, 10.25, 10.25, 
10.251, 10.251, 10.251), Col2 = c(10, 10.111, 10.222, 10.333, 
10.444, 10.555)), class = "data.frame", row.names = c("1", "2", 
"3", "4", "5", "6"))
Community
  • 1
  • 1
akrun
  • 874,273
  • 37
  • 540
  • 662
1

I'm not an xts officionado, but I think the first step is to convert those two columns into a POSIXt object.

as.POSIXct(dat$DATE, tz = "UTC", format = "%Y-%m-%d") + dat$TIME
# [1] "1993-01-04 09:35:38 UTC" "1994-01-05 09:35:41 UTC"
# [3] "1997-03-16 09:35:46 UTC" "2017-11-10 09:36:01 UTC"
# [5] "2001-08-28 09:36:05 UTC" "2006-04-20 09:40:07 UTC"

(BTW: assuming a "%H.%M.%S" format, I don't think you have have 36 minutes 61 seconds ... )

r2evans
  • 141,215
  • 6
  • 77
  • 149
  • 1
    Yeah, thanks. These aren't actual data I'm working with, but just number placeholders for the question. I got rid of the entries that didn't make sense. – D.J. P. May 27 '20 at 18:18