split data based on cumulative value of column in r - r

Question

I have following type of data:
myd <- data.frame (group = c(rep(1, 15), rep(2, 15)),
distance = c(0, 4, 8,9,11, 14,18,19,23, 24, 29,30,35,40, 43,
0, 8,9,9,12, 13,14,15,16, 18, 23,24,28, 29, 30),
var1 = c(1:15, 11:25), var2 = 1:30, var3 = 1:30)
myd
group distance var1 var2 var3
1 1 0 1 1 1
2 1 4 2 2 2
3 1 8 3 3 3
4 1 9 4 4 4
5 1 11 5 5 5
6 1 14 6 6 6
7 1 18 7 7 7
8 1 19 8 8 8
9 1 23 9 9 9
10 1 24 10 10 10
11 1 29 11 11 11
12 1 30 12 12 12
13 1 35 13 13 13
14 1 40 14 14 14
15 1 43 15 15 15
16 2 0 11 16 16
17 2 8 12 17 17
18 2 9 13 18 18
19 2 9 14 19 19
20 2 12 15 20 20
21 2 13 16 21 21
22 2 14 17 22 22
23 2 15 18 23 23
24 2 16 19 24 24
25 2 18 20 25 25
26 2 23 21 26 26
27 2 24 22 27 27
28 2 28 23 28 28
29 2 29 24 29 29
30 2 30 25 30 30
I have multiple group levels (than 2 shown above). For each distance (say mile posts in highway) starts with 0 and is cumulative to end to the group. I want to split the data (make bins) in such way that each group is approximately of distance 10. The resulting split data will be like:
data group1subset1
group distance var1 var2 var3
1 1 0 1 1 1
2 1 4 2 2 2
3 1 8 3 3 3
4 1 9 4 4 4
data group1subset2
5 1 11 5 5 5
6 1 14 6 6 6
7 1 18 7 7 7
8 1 19 8 8 8
data group1subset3
9 1 23 9 9 9
10 1 24 10 10 10
11 1 29 11 11 11
12 1 30 12 12 12
data group1subset4
13 1 35 13 13 13
14 1 40 14 14 14
data group1subset5
15 1 43 15 15 15
=====
data group2subset1
16 2 0 11 16 16
17 2 8 12 17 17
18 2 9 13 18 18
19 2 9 14 19 19
data group2subset2
20 2 12 15 20 20
21 2 13 16 21 21
22 2 14 17 22 22
23 2 15 18 23 23
24 2 16 19 24 24
25 2 18 20 25 25
data group2subset3
26 2 23 21 26 26
27 2 24 22 27 27
28 2 28 23 28 28
29 2 29 24 29 29
30 2 30 25 30 30
I need to automize the process as the real data is really big. Please suggest how can we do it?

Answer 1

I'd actually use cut to accomplish this:
maxd <- (max(myd$distance) %/% 10 * 10) + 10
transform(myd,cutdist = cut(distance, breaks = seq(0,maxd, by = 10),
include.lowest = TRUE))
group distance var1 var2 var3 cumdist cutdist
1 1 0 1 1 1 0 [0,10]
2 1 4 2 2 2 4 [0,10]
3 1 8 3 3 3 12 [0,10]
4 1 9 4 4 4 21 [0,10]
5 1 11 5 5 5 32 (10,20]
6 1 14 6 6 6 46 (10,20]
7 1 18 7 7 7 64 (10,20]
8 1 19 8 8 8 83 (10,20]
9 1 23 9 9 9 106 (20,30]
10 1 24 10 10 10 130 (20,30]
11 1 29 11 11 11 159 (20,30]
12 1 30 12 12 12 189 (20,30]
13 1 35 13 13 13 224 (30,40]
14 1 40 14 14 14 264 (30,40]
15 1 43 15 15 15 307 (40,50]
16 2 0 11 16 16 307 [0,10]
17 2 8 12 17 17 315 [0,10]
18 2 9 13 18 18 324 [0,10]
19 2 9 14 19 19 333 [0,10]
20 2 12 15 20 20 345 (10,20]
21 2 13 16 21 21 358 (10,20]
22 2 14 17 22 22 372 (10,20]
23 2 15 18 23 23 387 (10,20]
24 2 16 19 24 24 403 (10,20]
25 2 18 20 25 25 421 (10,20]
26 2 23 21 26 26 444 (20,30]
27 2 24 22 27 27 468 (20,30]
28 2 28 23 28 28 496 (20,30]
29 2 29 24 29 29 525 (20,30]
30 2 30 25 30 30 555 (20,30]
There's actually no need to calculate cumulative distance, since you want to keep them in groups of multiples of 10

Answer 2

What about
split(myd, interaction(myd$distance %/%10, myd$group, drop=F), drop=T)
here we bin the distance based on how many times it's divisible by 10 and create an interaction with group to make distinct categories for each. We us that as a factor for split(). This puts 0 with the first group and 10 with the second so some of your example breaks are slightly different with the code, but it seems comparable.