How to create a sequence column based on sequences' starts and ends - r

Question

I've got a two columns that contain information about sequences' starts and ends. I want to create a sequence column from that, i.e. each sequence starts when a seq_start is 1 and ends in first row appearing after seq_start = 1 in which seq_end = 1. How can I do it with tidyverse? The data is shown below, where seq is expected output. Please note that when seq_end = 1 and seq_start = 1 within the same rows this produces the sequence of length one.
structure(list(seq_start = c(NA, NA, NA, NA, NA, 1, NA, NA, NA,
NA, NA, 1, NA, 1, NA, NA, NA, NA, NA, NA, 1, 1, NA, NA, NA, NA,
NA, 1, 1, NA, NA, 1, NA, NA, NA, 1, NA, NA, NA, NA, NA, NA, NA,
NA, NA, NA, NA, NA, NA, 1, NA, NA, NA, NA, NA, NA, NA, NA, 1,
NA), seq_end = c(NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, 1L,
1L, 1L, 1L, NA, NA, 1L, 1L, 1L, NA, 1L, NA, NA, NA, NA, NA, 1L,
1L, NA, NA, 1L, 1L, NA, 1L, 1L, 1L, 1L, NA, NA, NA, 1L, 1L, NA,
NA, NA, NA, NA, NA, 1L, NA, 1L, 1L, NA, 1L, 1L, NA, NA, 1L, 1L,
1L), seq = c(NA, NA, NA, NA, NA, 1L, 1L, 1L, 1L, 1L, 1L, 2L,
NA, 3L, NA, NA, NA, NA, NA, NA, 4L, 5L, 5L, 5L, 5L, 5L, 5L, 6L,
7L, 7L, 7L, 8L, NA, NA, NA, 9L, NA, NA, NA, NA, NA, NA, NA, NA,
NA, NA, NA, NA, NA, 10L, 10L, NA, NA, NA, NA, NA, NA, NA, 11L,
NA)), .Names = c("seq_start", "seq_end", "seq"), class = c("tbl_df",
"tbl", "data.frame"), row.names = c(NA, -60L))

Answer 1

Here's a solution that makes heavy use of dplyr package's lag() function, along with cumsum() from the base package, to produce the expected result. It's probably not the most succinct solution out there, but I do think it's reasonably intuitive to understand:
d <- d %>%
# new.seq.starts starts from 0, and increments by 1 every time seq_starts takes on
# the value 1, like this: 0, 0, 0, 1, 1, 1, 1, 2, 2, ...
# Rows with the same new.seq.starts value are thus part of the same "run".
mutate(new.seq.starts = cumsum(!is.na(seq_start))) %>%
# group by each "run"
group_by(new.seq.starts) %>%
# any.ending.so.far counts whether there has been ANY seq_end == 1 within the run yet.
# first.ending is TRUE only if it's the first row (within the run) to have an ending.
mutate(any.ending.so.far = cumsum(!is.na(seq_end)),
first.ending = any.ending.so.far == 1 &
(is.na(lag(any.ending.so.far)) | lag(any.ending.so.far) < 1)) %>%
ungroup() %>%
# result keeps the new.seq.starts values only if there's no ending yet (i.e.
# any.ending.so.far == 0), or only just ended (first.ending == TRUE). Otherwise,
# it takes on the value NA.
mutate(result = ifelse(new.seq.starts > 0 &
(any.ending.so.far == 0 | first.ending),
new.seq.starts, NA)) %>%
# Remove helper variables as they are no longer needed.
select(-c(new.seq.starts, any.ending.so.far, first.ending))
> all.equal(d$seq, d$result)
[1] TRUE