CODEWITHSUNDEEP
rdatetime
user1702490
user1702490

Reputation: 467

calculating time difference in R

I have a data with more that 3 million records having start.time and end.time as two of the variables. The first 10 obs are as follows:

   start.date start.time   end.date end.time
1  2012-07-13   15:01:32 2012-07-13 15:02:42
2  2012-07-05   18:26:31 2012-07-05 18:27:19
3  2012-07-14   20:23:21 2012-07-14 20:24:11
4  2012-07-29   16:09:54 2012-07-29 16:10:48
5  2012-07-21   14:58:32 2012-07-21 15:00:17
6  2012-07-04   15:36:31 2012-07-04 15:37:11
7  2012-07-22   18:28:31 2012-07-22 18:28:50
8  2012-07-09   21:08:42 2012-07-09 21:09:02
9  2012-07-05   09:44:52 2012-07-05 09:45:05
10 2012-07-02   18:50:47 2012-07-02 18:51:38

I need to calculate the difference between start.time and end.time.

I used the following code:

mbehave11$diff.time <- difftime(mbehave11$end.time, mbehave11$start.time, units="secs")

But I am getting this error:

Error in as.POSIXlt.character(x, tz, ...) : 
  character string is not in a standard unambiguous format
In addition: Warning messages:
1: In is.na.POSIXlt(strptime(xx, f <- "%Y-%m-%d %H:%M:%OS", tz = tz)) :
  Reached total allocation of 1535Mb: see help(memory.size)

Upvotes: 30

Views: 115426

Answers (2)

Dirk is no longer here
Dirk is no longer here

Reputation: 368181

You must turn your strings into date objects before you can do date/time arithmetic. Try this:

a) Reading your data:

R> dat <- read.table(textConnection("start.date start.time end.date end.time
2012-07-13   15:01:32 2012-07-13 15:02:42
2012-07-05   18:26:31 2012-07-05 18:27:19 
2012-07-14   20:23:21 2012-07-14 20:24:11"), header=TRUE)

b) Working on one observation:

 R>  strptime( paste(dat[,1], dat[,2]), "%Y-%m-%d %H:%M:%S")
 [1] "2012-07-13 15:01:32" "2012-07-05 18:26:31" "2012-07-14 20:23:21"

c) Working on the set, converting to numeric:

 R> as.numeric(difftime(strptime(paste(dat[,1],dat[,2]),"%Y-%m-%d %H:%M:%S"),
                        strptime(paste(dat[,3],dat[,4]),"%Y-%m-%d %H:%M:%S"))) 
 [1] -70 -48 -50
 R>

Edit Some seven years later by someone else below.

d) Just to explain the results -70 -48 -50 above take a look at the example row by row:

[2012-07-13 15:01:32] - [2012-07-13 15:02:42] = -70 seconds,  
[2012-07-05 18:26:31] - [2012-07-05 18:27:19] = -48 seconds,  
[2012-07-14 20:23:21] - [2012-07-14 20:24:11] = -50 seconds

Upvotes: 47

cloudscomputes
cloudscomputes

Reputation: 1474

I think you can use the lubridate package
it has a method called ymd_hms
you can use that to get the time from string: it is much faster for large data set

library(lubridate)
dat <- read.table(textConnection("start.date start.time end.date end.time
2012-07-13   15:01:32 2012-07-13 15:02:42
2012-07-05   18:26:31 2012-07-05 18:27:19 
2012-07-14   20:23:21 2012-07-14 20:24:11"), header=TRUE)
starttime = ymd_hms(paste(dat[,1], dat[,2]))
endtime = ymd_hms(paste(dat[,3], dat[,4]))
interval = difftime(endtime,starttime,units = "secs")

or you can just do it in one line, but it takes longer time for big dataset:

difftime(paste(dat[,3], dat[,4]),paste(dat[,1], dat[,2]),units = "secs")

Upvotes: 6

Related Questions