Question

So I’ve looked at this and this, and can’t seem to find a good example of a way to loop or iterate over a factor (or are their other / better ways to do this?). I have a data frame for which I have:

> head(frame)
    x1     x2     DateTime
  1 100   5   2010-06-01 05:32:46
  2 105   3   2010-06-01 05:32:23
  3 47    20  2010-06-01 05:32:34
  4 56    6   2010-06-01 05:33:16
  5 98    11  2010-06-01 05:54:12
  6 84    9   2010-06-01 05:54:05

and I can create a factor based on time like so: fact <- cut(frame$DateTime, "1 hour") from there, how would I go about extracting the first and last elements of frame$x2 given the factor I’ve created? (or for that matter, the nth element of the cut).

would it be something like:

test <- split(frame$x2, fact)

Answer 1

I’m not sure what you mean by “extract” the nth element (i.e. do you want it in the same object, a new object, do you want something the same length as the original object, etc.).

I would use ave because it operates on factor groups. Note that this assumes that your data.frame is sorted by frame$DateTime.

frame <- structure(list(
  x1 = c(100L, 105L, 47L, 56L, 98L, 84L),
  x2 = c(5L, 3L, 20L, 6L, 11L, 9L),
  DateTime = structure(c(1275388366, 1275388343, 1275388354, 1275388396,
    1275389652, 1275389645), class = c("POSIXct", "POSIXt"), tzone = ""),
  fact = structure(c(1L, 1L, 1L, 1L, 1L, 1L),
    .Label = "2010-06-01 05:00:00", class = "factor")),
    .Names = c("x1", "x2", "DateTime", "fact"),
  row.names = c(NA, -6L), class = "data.frame")

transform(frame, firstx2=ave(x2, fact, FUN=function(x) x[1]),
                 lastx2 =ave(x2, fact, FUN=function(x) x[length(x)]))

# These lines do the same as the `transform` line(s)
frame$firstx2 <- ave(frame$x2, frame$fact, FUN=function(x) x[1])
frame$lastx2  <- ave(frame$x2, frame$fact, FUN=function(x) x[length(x)])