CODEWITHSUNDEEP
r
Madhu Sareen
Madhu Sareen

Reputation: 549

Random incremental timestamp in R

I want to generate a series of 500 TimeStamp starting Jan 1st, 2016 in such a way that the increment of time stamp should look like something as below.

Sample:

TimeStamp
2016-01-01 00:00:01
2016-01-01 00:00:12
2016-01-01 00:00:15
2016-01-01 00:01:23
2016-01-01 00:02:31
2016-01-01 00:02:38
2016-01-01 00:03:48
2016-01-01 00:03:55
.....

What I am doing as of now is:

SampleData <-  as.data.frame(list(Var1=1:500, Var2=rnorm(1, 500, 500)))

rDate <- function(sDate, eDate, SampleData){   
  lenDate <- dim(sampledata)[1]    
  seqDays <- seq.Date(as.Date(sDate), as.Date(eDate), by="day")  
  aDay <- runif(lenDate, 1, length(seqDays))  
  Date <- seqDays[aDay]  
}

SampleData$TimeStamp <- rDate("2016-01-01", "2016-12-31", SampleData)
SampleData <- SampleData[order(SampleData$TimeStamp), ]
row.names(SampleData) <- NULL
head(SampleData)

But this will produce the following result:

  Var1     Var2  TimeStamp
1  200 1020.469 2016-01-01
2  100 1020.469 2016-01-02
3  344 1020.469 2016-01-02
4  447 1020.469 2016-01-04
5  453 1020.469 2016-01-05
6  478 1020.469 2016-01-05

Which is not what I wanted.

Could someone please help?

Upvotes: 1

Views: 826

Answers (3)

Ronak Shah
Ronak Shah

Reputation: 389047

Maybe something like, 

as.POSIXct("2016-01-01 00:00:00") + sort(sample(1:1000, 500))

We can check this for 5 samples

as.POSIXct("2016-01-01 00:00:00") + sort(sample(1:1000, 5))

#[1] "2016-01-01 00:01:53 IST" "2016-01-01 00:02:06 IST" "2016-01-01 00:03:19 IST"
#[4] "2016-01-01 00:07:31 IST" "2016-01-01 00:12:26 IST"

This will add randomly 1 to 1000 seconds in an incremental fashion in 1st of Jan 2016. To further increase the range we can increase the sequence from 1:1000 to any number you wish.

Another solution which takes advantage of entire range is 

startTime <- as.POSIXct("2016-01-01")
endTime <- as.POSIXct("2016-12-31")
sample(seq(startTime, endTime, 1), 500)

Here we generate sequence for every second from our startTime to our endTime and then take random 500 values from it. Although, this is complete but this would become slow as the difference between startTime and endTime increases.

Upvotes: 1

Manoj Kumar
Manoj Kumar

Reputation: 5647

Here I got something for you...

RandomTimeStamp <- function(M, sDate="2016/01/01", eDate="2016/12/31") {
sDate <- as.POSIXct(as.Date(sDate))
eDate <- as.POSIXct(as.Date(eDate))
dTime <- as.numeric(difftime(eDate, sDate, unit="sec"))
sTimeStamp <- sort(runif(M, 0, dTime))
TimeStamp <- sDate + sTimeStamp
}

print(RandomTimeStamp(500))

This produces the result as:

[1] "2012-01-01 18:26:53 IST" "2012-01-02 11:35:47 IST" "2012-01-02 15:02:23 IST" "2012-01-02 19:19:25 IST"
  [5] "2012-01-03 04:48:13 IST" "2012-01-03 21:05:42 IST" "2012-01-03 21:16:06 IST" "2012-01-04 21:05:08 IST"
  [9] "2012-01-05 05:47:13 IST" "2012-01-05 06:27:44 IST" "2012-01-05 06:40:42 IST" "2012-01-05 21:56:45 IST"
 [13] "2012-01-06 22:36:40 IST" "2012-01-07 03:48:37 IST" "2012-01-07 12:55:25 IST" "2012-01-07 20:52:19 IST" .........

You might want to tweak around the code... :)

Upvotes: 1

GGamba
GGamba

Reputation: 13680

Just change seq.Date into seq.POSIXt and as.Date into as.POSIXct

SampleData <-  as.data.frame(list(Var1=1:500, Var2=rnorm(1, 500, 500)))

rDate <- function(sDate, eDate, SampleData){   
    lenDate <- dim(SampleData)[1]    
    seqDays <- seq.POSIXt(as.POSIXct(sDate), as.POSIXct(eDate), by="secs")  
    aDay <- runif(lenDate, 1, length(seqDays))  
    Date <- seqDays[aDay]  
}

SampleData$TimeStamp <- rDate("2016-01-01", "2016-12-31", SampleData)
SampleData <- SampleData[order(SampleData$TimeStamp), ]
row.names(SampleData) <- NULL
head(SampleData)

  Var1     Var2           TimeStamp
1   29 660.4593 2016-01-01 13:25:31
2  213 660.4593 2016-01-02 07:17:10
3  115 660.4593 2016-01-05 01:07:48
4  358 660.4593 2016-01-05 06:24:41
5  276 660.4593 2016-01-06 10:02:18
6   49 660.4593 2016-01-06 21:56:25

Upvotes: 1

Related Questions