Using dplyr to filter all rows from one variable until another

Question

My data looks like this:

data <- data.frame(
  value = runif(10)
  id = c("junk","start","1","2","end","morejunk","junk","start","4","end")
)

I want to use filter() to extract everything from the id "start" until the id "end" . The problem is the number of observations between the starting row and ending row vary, so I can't filter every x rows. Is there a way to use filter() in a way that I could specify from = "start" until = "end" ?

Answer 1

You can first identify where "start" and "end" are. Then using those pairwise of indices to index the data.frame. This assumes that there is a corresponding pair of start and end each time.

set.seed(0L)
data <- data.frame(
    value = runif(10),
    id = c("junk","start","1","2","end","morejunk","junk","start","4","end")
)
idx <- which(data$id %in% c("start", "end"))
lapply(split(idx, ceiling(seq_along(idx)/2)), function(x) data[x[1]:x[2],])

Answer 2

You can

• use which to identify row indices with "start" and "end" ,
• add and subtract 1 from each, respectively so as not to include those rows,
• pass those sequences to : in parallel via Map ,
• unlist to simplify the list to a vector, and
• subset with slice

which leaves

library(dplyr)
set.seed(47)

data <- data.frame(
    value = runif(10),
    id = c("junk","start","1","2","end","morejunk","junk","start","4","end")
)

data %>% slice(unlist(Map(`:`, 
                          which(.$id == 'start') + 1, 
                          which(.$id == 'end') - 1)))
#> # A tibble: 3 × 2
#>       value     id
#>       <dbl> <fctr>
#> 1 0.7615020      1
#> 2 0.8224916      2
#> 3 0.5433097      4

or in base,

data[unlist(Map(`:`, 
                which(data$id == 'start') + 1, 
                which(data$id == 'end') - 1)), ]
#>       value id
#> 3 0.7615020  1
#> 4 0.8224916  2
#> 9 0.5433097  4