group elements in a list of numbers according to a threshold in R

Question

suppose i have a list of number:

numbers = c(1, 2, 5, 6, 3, 2, 3, 7, 6)

and a threshold

threshold = 6

I want to group the elements in the number list according to the threshold -- summing up the elements until you reach/over the threshold. Save the index of the first element of each group, so the desired output is:

1, 4, 5, 8, 9

1 because it's the start

4 because 1+2+5 = 8 > 6 and this include 3 elements

it is similar to this post Sum list of numbers until threshold but we want to keep adding until we reach the end of the list.

Does this answer your question? [Conditional cumsum with reset](https://stackoverflow.com/questions/34531568/conditional-cumsum-with-reset) — AnilGoyal, Mar 26 '21 at 08:29

score 2 · Accepted Answer · answered Mar 26 '21 at 08:03

2

You can use the function cumsumbinning from the package MESS :

numbers = c(1, 2, 5, 6, 3, 2, 3, 7, 6)
threshold = 6
which(!duplicated(MESS::cumsumbinning(numbers, threshold - 1, cutwhenpassed=TRUE), fromLast = TRUE))
#[1] 3 4 7 8 9

If you want the first element in the group.

which(!duplicated(MESS::cumsumbinning(numbers, threshold - 1, cutwhenpassed=TRUE)))
#[1] 1 4 5 8 9

cumsumbinning with cutwhenpassed=TRUE creates a new group whenever the value passes the threshold value hence I used threshold - 1.

MESS::cumsumbinning(numbers, threshold - 1, cutwhenpassed=TRUE)
#[1] 1 1 1 2 3 3 3 4 5

answered Mar 26 '21 at 08:03

Ronak Shah

377,200
20
156
213

is there another way to do this without the MESS packages? I keep getting errors when downloading the MESS packages. – Math Avengers Mar 26 '21 at 14:09
Another way I can think of is messy loops. What is the issue that you are facing downloading the package? I installed it myself while trying to answer this question and I could download it without any problem using `install.packages('MESS')` – Ronak Shah Mar 27 '21 at 00:22
I am getting lots of errors like `ERROR: dependencies 'glmnet', 'kinship2' are not available for package 'MESS'` – Math Avengers Mar 29 '21 at 02:02
I think the main problem is this line `cannot find -lgfortran` – Math Avengers Mar 29 '21 at 02:07

score 1 · Answer 2 · answered Mar 26 '21 at 08:40

I think this will also serve the purpose

library(purrr)
setdiff(which(accumulate(numbers, ~ifelse(.x + .y <  6, .y + .x, 0)) ==0), which(numbers == 0))

#check on new vector numbers
numbers = c(1, 2, 5, 6, 3, 2, 3, 7, 6, 0, 2, 3, 7, 1, 2)

setdiff(which(accumulate(numbers, ~ifelse(.x + .y <  6, .y + .x, 0)) ==0), which(numbers == 0))
[1]  3  4  7  8  9 13

group elements in a list of numbers according to a threshold in R

2 Answers2