data.table::rleid for dplyr

[ghost]

Hi

it would be cool to have data.table::rleid in dplyr for grouping, window functions etc.

http://stackoverflow.com/questions/33507868/is-there-a-dplyr-equivalent-to-data-tablerleid

myrleid <- function(x) {
x <- rle(x)$lengths
rep(seq_along(x), times=x)
}

DT <- data.frame(grp=rep(c("A", "B", "C", "A", "B"), c(2,2,3,1,2)), value=1:10)
DT %<>% mutate(rlid = myrleid(grp))
DT

Greetings
Christof

[hadley]

group_indices() ?

[yutannihilation]

Just for future reference, group_indices() is NOT equivalent to data.table::rleid().

For example, suppose we want to treat first three "a" and last "a" in the data bellow are different groups:

d <- data.frame(x = c("a", "a", "a", "b", "b", "a"), stringsAsFactors = FALSE)

dplyr::group_indices(d, x)
#> [1] 1 1 1 2 2 1

data.table::rleid(d$x)
#> [1] 1 1 1 2 2 3

This can be implemented easily like this:

my_rleid <- function(x) {
  1L + cumsum(dplyr::coalesce(x != dplyr::lag(x), FALSE))
}

(But, I think this is not dplyr's job. Maybe tidyr can fit since this is the similar problem which functions like fill() try to solve in that they both are about how to work with consecutive rows.)