Count consecutive strings of zeroes and ones over multiple groups - r

Question

There have been several discussions about counting consecutive strings of zeroes and ones (or other values) using functions like rle or cumsum. I have played around with these functions, but I can't easily figure out how to get them to apply to my specific problem.
I am working with ecological presence/absence data ("pres.abs" = 1 or 0) organized by time ("year") and location ("id"). For each location id, I would like to separately calculate the length of consecutive ones and zeroes through time. Where these cannot be calculated, I want to return "NA".
Below is a sample of what the data looks like (first 3 columns) and the output I am hoping to achieve (last 2 columns). Ideally, this would be a pretty fast function avoiding for-loops since the real data frame contains ~15,000 rows.
year = rep(1:10, times=3)
id = c(rep(1, times=10), rep(2, times=10), rep(3, times=10))
pres.abs.id.1 = c(0, 0, 0, 1, 1, 1, 0, 0, 1, 1) #Pres/abs data at site 1 across time
pres.abs.id.2 = c(1, 1, 0, 1, 0, 0, 1, 0, 0, 0) #Pres/abs data at site 2 across time
pres.abs.id.3 = c(0, 0, 0, 0, 0, 1, 1, 1, 1, 1) #Pres/abs data at site 3 across time
pres.abs = c(pres.abs.id.1, pres.abs.id.2, pres.abs.id.3)
dat = data.frame(id, year, pres.abs)
dat$cumul.zeroes = c(1,2,3,NA,NA,NA,1,2,NA,NA,NA,NA,1,NA,1,2,NA,1,2,3,1,2,3,4,5,NA,NA,NA,NA,NA)
dat$cumul.ones = c(NA,NA,NA,1,2,3,NA,NA,1,2,1,2,NA,1,NA,NA,1,NA,NA,NA,NA,NA,NA,NA,NA,1,2,3,4,5)
> dat
id year pres.abs cumul.zeroes cumul.ones
1 1 1 0 1 NA
2 1 2 0 2 NA
3 1 3 0 3 NA
4 1 4 1 NA 1
5 1 5 1 NA 2
6 1 6 1 NA 3
7 1 7 0 1 NA
8 1 8 0 2 NA
9 1 9 1 NA 1
10 1 10 1 NA 2
11 2 1 1 NA 1
12 2 2 1 NA 2
13 2 3 0 1 NA
14 2 4 1 NA 1
15 2 5 0 1 NA
16 2 6 0 2 NA
17 2 7 1 NA 1
18 2 8 0 1 NA
19 2 9 0 2 NA
20 2 10 0 3 NA
21 3 1 0 1 NA
22 3 2 0 2 NA
23 3 3 0 3 NA
24 3 4 0 4 NA
25 3 5 0 5 NA
26 3 6 1 NA 1
27 3 7 1 NA 2
28 3 8 1 NA 3
29 3 9 1 NA 4
30 3 10 1 NA 5
Thanks very much for your help.

Answer 1

Here's a base R way using rle and sequence:
dat <- within(dat, {
cumul.counts <- unlist(lapply(split(pres.abs, id), function(x) sequence(rle(x)$lengths)))
cumul.zeroes <- replace(cumul.counts, pres.abs == 1, NA)
cumul.ones <- replace(cumul.counts, pres.abs == 0, NA)
rm(cumul.counts)
})
# id year pres.abs cumul.ones cumul.zeroes
# 1 1 1 0 NA 1
# 2 1 2 0 NA 2
# 3 1 3 0 NA 3
# 4 1 4 1 1 NA
# 5 1 5 1 2 NA
# 6 1 6 1 3 NA
# 7 1 7 0 NA 1
# 8 1 8 0 NA 2
# 9 1 9 1 1 NA
# 10 1 10 1 2 NA
# 11 2 1 1 1 NA
# 12 2 2 1 2 NA
# 13 2 3 0 NA 1
# 14 2 4 1 1 NA
# 15 2 5 0 NA 1
# 16 2 6 0 NA 2
# 17 2 7 1 1 NA
# 18 2 8 0 NA 1
# 19 2 9 0 NA 2
# 20 2 10 0 NA 3
# 21 3 1 0 NA 1
# 22 3 2 0 NA 2
# 23 3 3 0 NA 3
# 24 3 4 0 NA 4
# 25 3 5 0 NA 5
# 26 3 6 1 1 NA
# 27 3 7 1 2 NA
# 28 3 8 1 3 NA
# 29 3 9 1 4 NA
# 30 3 10 1 5 NA

Answer 2

Here's one option with dplyr:
require(dplyr)
dat %>%
group_by(id, x = cumsum(c(0,diff(pres.abs)) != 0)) %>%
mutate(cumul.zeros = ifelse(pres.abs, NA_integer_, row_number()),
cumul.ones = ifelse(!pres.abs, NA_integer_, row_number())) %>%
ungroup() %>% select(-x)
#Source: local data frame [30 x 5]
#
# id year pres.abs cumul.zeros cumul.ones
#1 1 1 0 1 NA
#2 1 2 0 2 NA
#3 1 3 0 3 NA
#4 1 4 1 NA 1
#5 1 5 1 NA 2
#6 1 6 1 NA 3
#7 1 7 0 1 NA
#8 1 8 0 2 NA
#9 1 9 1 NA 1
#10 1 10 1 NA 2
#11 2 1 1 NA 1
#12 2 2 1 NA 2
#13 2 3 0 1 NA
#14 2 4 1 NA 1
#15 2 5 0 1 NA
#16 2 6 0 2 NA
#17 2 7 1 NA 1
#18 2 8 0 1 NA
#19 2 9 0 2 NA
#20 2 10 0 3 NA
#21 3 1 0 1 NA
#22 3 2 0 2 NA
#23 3 3 0 3 NA
#24 3 4 0 4 NA
#25 3 5 0 5 NA
#26 3 6 1 NA 1
#27 3 7 1 NA 2
#28 3 8 1 NA 3
#29 3 9 1 NA 4
#30 3 10 1 NA 5