Subsampling without replacement in R [duplicate] - r

Question

This question already has an answer here:
Sample without replacement, or duplicates, in R
(1 answer)
Closed 1 year ago.
There's a very similar question to mine on stack, but that doesn't directly answer my question.
I have abundance data for 250 species across 1000 sites. Species are columns, sites are rows. My abundances data look something like the data in the linked post above.
0 0 3 0 0 201 0 0 0 82
0 23 5 0 0 0 0 0 0 0
9 0 0 0 0 12 0 0 0 913
0 7 91 0 8 0 0 92 9 0
131 12 0 410 0 0 0 3 0 0
If I wanted to sample 50 individuals from each site, without replacement, how can I do this?
Focusing on code for single sites for now.
This code:
samples <- sample(1:ncol(abundances), 50, rep=FALSE, prob=abundances[1,]) doesn't work unless I change to rep=TRUE. However, I need sampling WITHOUT replacement.
I don't want to use sample(abundances[1,], 50, rep=FALSE) because then instead of sampling individuals, it samples species and will report the whole value in that row (i.e. species 6 may occur 201 times at site 1, it'll report 201, rather than 1 individual from that species, resulting in >50 individuals in final subsample).
I essentially want an output identical to what user Dinre answered in post above, but without it being for bootstrapping. I just want to sample without replacement. This process will ultimately be integrated into a for loop for a subsample from each site.

Answer 1

Here is a way to sample vector elements from each row with the sum of each sampled row equal to a chosen integer, the size. In the code below, n <- 5 is passed as function argument size. The call to runif adds an element of randomness to the sampling function.
fun <- function(x, size){
x <- x*runif(length(x))
y <- size*x/sum(x)
round(y)
}
set.seed(2021)
n <- 5
t(apply(df1, 1, fun, size = n))
# V1 V2 V3 V4 V5 V6 V7 V8 V9 V10
#[1,] 0 0 0 0 0 3 0 0 0 2
#[2,] 0 4 1 0 0 0 0 0 0 0
#[3,] 0 0 0 0 0 0 0 0 0 5
#[4,] 0 0 3 0 0 0 0 1 0 0
#[5,] 0 0 0 5 0 0 0 0 0 0
Data
Here is the question's data in dput format.
X <-
structure(c(0L, 0L, 9L, 0L, 131L, 0L, 23L, 0L, 7L, 12L, 3L, 5L,
0L, 91L, 0L, 0L, 0L, 0L, 0L, 410L, 0L, 0L, 0L, 8L, 0L, 201L,
0L, 12L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 92L, 3L, 0L,
0L, 0L, 9L, 0L, 82L, 0L, 913L, 0L, 0L), .Dim = c(5L, 10L), .Dimnames = list(
NULL, c("V1", "V2", "V3", "V4", "V5", "V6", "V7", "V8", "V9", "V10")))