1

I want to use one column which indicates the start point for each sample and then flag the points (rows) that follow the start point up until a maximum amount of time is reached.

For example - my data (d) looks like:

> head(d)
  Sample Seconds Value FLAG
1      A     356     1    1
2      A     357     1   NA
3      A     358     9   NA
4      A     359     4   NA
5      A     400     1   NA
6      A     401     3   NA

A reproducible copy is here:

d <- structure(list(Sample = structure(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 
1L, 2L, 2L, 2L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L
), .Label = c("A", "B", "C"), class = "factor"), Seconds = c(356L, 
357L, 358L, 359L, 400L, 401L, 402L, 403L, 2955L, 2957L, 2959L, 
3001L, 3002L, 3004L, 2548L, 2549L, 2552L, 2553L, 2554L, 2555L, 
2556L, 2557L, 2558L), Value = c(1L, 1L, 9L, 4L, 1L, 3L, 7L, 2L, 
25L, 17L, 23L, 47L, 34L, 15L, 30L, 16L, 17L, 12L, 6L, 8L, 6L, 
6L, 5L), FLAG = c(1L, NA, NA, NA, NA, NA, NA, NA, 1L, NA, NA, 
NA, NA, NA, 1L, NA, NA, NA, NA, NA, NA, NA, NA)), .Names = c("Sample", 
"Seconds", "Value", "FLAG"), class = "data.frame", row.names = c(NA, 
-23L))

I only want the first five seconds of data for each sample. The flag indicates the first row of the sample (keep in mind, this is a simplified version - my real data requires that I set up a flag handle to find start points). I want to grab the row with the start point (Flag=1) and then go to d$Seconds and put a 1 value in each row within a 5 second window from the start point. I can't just add five because some samples have a point at 4 and then 6 seconds from the start. I am working with a large dataset so I am also trying to avoid a for loop. Any ideas? (sorry for the data format - I haven't posted before and it wouldn't let me put an image)

joran
  • 169,992
  • 32
  • 429
  • 468
user1508031
  • 11
  • 1
  • FYI, see [this](http://stackoverflow.com/q/5963269/324364) question for how to make your example reproducible, as I have done for you. In general, how your data looks when you print it out on the screen is virtually useless for those wishing to help. Even _more_ useless would be an actual image or screenshot. – joran Jul 07 '12 at 00:40

1 Answers1

1

Here's one method in base R using by:

d$within5 <- unlist(
                by(
                    d,
                    d$Sample,
                    function (x) x$Seconds <= (x$Seconds[!is.na(x$FLAG)]+5)
                  )
                   )

Result:

> head(d)
  Sample Seconds Value FLAG within5
1      A     356     1    1    TRUE
2      A     357     1   NA    TRUE
3      A     358     9   NA    TRUE
4      A     359     4   NA    TRUE
5      A     400     1   NA   FALSE
6      A     401     3   NA   FALSE
thelatemail
  • 91,185
  • 12
  • 128
  • 188