dplyr: Find first non-zero element and last non-zero element and trim vector, by group

Question

I have a similar question to this post (How to find first non-zero element and last non-zero element and TRIM vector)

but I am looking for a dplyr solution, as I want to include a

 d %>% 
  group_by(id) %>%

statement in addition.

I have a data frame:

d<-data.frame(time = factor(c("00:00","00:15","00:30","00:45", "01:00","01:15","01:30","01:45","02:00","02:40" )), q=c(0,0,100,0,0,100,0,0,0,0),p=c(.25,.25,.25,.25,.25,.25,.25,.25,.25,.25))
d
    time   q    p
1  00:00   0 0.25
2  00:15   0 0.25
3  00:30 100 0.25
4  00:45   0 0.25
5  01:00   0 0.25
6  01:15 100 0.25
7  01:30   0 0.25
8  01:45   0 0.25
9  02:00   0 0.25
10 02:40   0 0.25

I would like to eliminate rows of the data frame that are BEFORE the first non-zero index of column "q" AND AFTER the last non-zero index of column "q". In the case above the results should look like this:

  00:30 100 0.25
  00:45   0 0.25
  01:00   0 0.25
  01:15 100 0.25

@akrun gave a solution to this question:

 indx <- which(d$q!=0)
 d[indx[1L]:indx[length(indx)],]

This works, but I am looking for a dplyr solution, as I want to perform this calculation across multiple groups.

Answer 1

One option could be:

d %>%
 filter(row_number() %in% Reduce(`:`, which(q != 0)))

   time   q    p
1 00:30 100 0.25
2 00:45   0 0.25
3 01:00   0 0.25
4 01:15 100 0.25

Answer 2

d<-data.frame(time = factor(c("00:00","00:15","00:30","00:45", "01:00","01:15","01:30","01:45","02:00","02:40" )), q=c(0,0,100,0,0,100,0,0,0,0),p=c(.25,.25,.25,.25,.25,.25,.25,.25,.25,.25))

library(dplyr, warn.conflicts = F)

d %>% filter(cumsum(q != 0) != 0) %>%
  filter(rev(cumsum(rev(q != 0))) != 0 )
#>    time   q    p
#> 1 00:30 100 0.25
#> 2 00:45   0 0.25
#> 3 01:00   0 0.25
#> 4 01:15 100 0.25

^{Created on 2021-06-12 by the reprex package (v2.0.0)}