CODEWITHSUNDEEP
rloopsfor-loopnested-loops
sedeh
sedeh

Reputation: 7313

R nested loop returning only the last iteration

I am trying to replicate 81-place cryoboxes used in lab storage system using a nested for loop. The following code illustrates the problem using 3-place boxes:

urine_random_df <- as.data.frame(c(seq(from = 10, to = 12, by = 1)))
boxcells <- vector()
cell_placeholder <- as.data.frame(c(seq(from = 1, to = 3, by = 1)))
for (i in 1: 3){
        #boxcells <- paste0("NEW", sprintf("%04d", as.numeric(urine_random_df[i,])))
        for (j in 1: nrow(cell_placeholder)){
                boxcells <- c(boxcells, paste(paste0("NEW", sprintf("%04d", as.numeric(urine_random_df[i,]))), cell_placeholder[j,], sep = "-"))        
        }

}


boxcells <- data.frame(boxcells)
names(boxcells) <- "box cells"
boxcells

The result of above is:

box cells
1 NEW0010-1
2 NEW0010-2
3 NEW0010-3
4 NEW0011-1
5 NEW0011-2
6 NEW0011-3
7 NEW0012-1
8 NEW0012-2
9 NEW0012-3

However, I would like to group the cells under their respective boxes like so:

   box cells
1  NEW0010
2  NEW0010-1
3  NEW0010-2
3  NEW0010-3
4  NEW0011
5  NEW0011-1
6  NEW0011-2
7  NEW0011-3
8  NEW0012
9  NEW0012-1
10 NEW0012-2
11 NEW0012-3

I tried to achieve this by adding boxcells <- paste0("NEW", sprintf("%04d", as.numeric(urine_random_df[i,]))) in the outer loop. When I re-ran the code with this piece, I get only the last box like so:

  box cells
1   NEW0012
2 NEW0012-1
3 NEW0012-2
4 NEW0012-3

It appears each iteration of the loop erases the last such that upon completion of the entire loop, only the last box remains. I found an existing thread here which suggests moving the "initialisation statements" outside the loop. However, in this case, the initialisation statements urine_random_df..., boxcells... and cell_placeholder... are already outside the loop. Thoughts?

Upvotes: 1

Views: 2047

Answers (3)

agstudy
agstudy

Reputation: 121568

If you want to group result better to choose another structure to store your result. Here 2 ways to simply resolve your problem: 

nn <- paste0('NEW',sprintf("%04d",10:12))

using a list

setNames(lapply(nn,function(x){
  paste(x,1:3,sep='-')
}),nn)

$NEW0010
[1] "NEW0010-1" "NEW0010-2" "NEW0010-3"

$NEW0011
[1] "NEW0011-1" "NEW0011-2" "NEW0011-3"

$NEW0012
[1] "NEW0012-1" "NEW0012-2" "NEW0012-3"

using data.frame

transform(expand.grid(nn,1:3),Var2=paste(Var1,Var2,sep='-'))

     Var1      Var2
1 NEW0010 NEW0010-1
2 NEW0011 NEW0011-1
3 NEW0012 NEW0012-1
4 NEW0010 NEW0010-2
5 NEW0011 NEW0011-2
6 NEW0012 NEW0012-2
7 NEW0010 NEW0010-3
8 NEW0011 NEW0011-3
9 NEW0012 NEW0012-3

Upvotes: 0

Ben
Ben

Reputation: 42283

I think the complication here comes from collecting the output of the loop as a vector rather than a list.

Here it is using a list, then unlisting and converting to a dataframe. The output is exactly as requested

urine_random_df <- as.data.frame(c(seq(from = 10, to = 12, by = 1)))
boxcells <- list()
cell_placeholder <- as.data.frame(c(seq(from = 1, to = 3, by = 1)))
n <- nrow(cell_placeholder)
for (i in 1:n){
  tmp <- vector()
tmp <- c(paste0("NEW", sprintf("%04d", as.numeric(urine_random_df[i,]))))
      for (j in 1:n){
        tmp <- c(tmp, paste(paste0("NEW", sprintf("%04d", as.numeric(urine_random_df[i,]))), cell_placeholder[j,], sep = "-"))        
      }
      boxcells[[i]] <- tmp 
}


boxcells <- data.frame(unlist(boxcells))
names(boxcells) <- "box cells"
boxcells

Which gives:

 box cells
1    NEW0010
2  NEW0010-1
3  NEW0010-2
4  NEW0010-3
5    NEW0011
6  NEW0011-1
7  NEW0011-2
8  NEW0011-3
9    NEW0012
10 NEW0012-1
11 NEW0012-2
12 NEW0012-3

Upvotes: 1

David Arenburg
David Arenburg

Reputation: 92282

I can think of very rare situations when you would do a nested for loop in R, even single for loop is very rare.

I would solve this by doing something like

temp <- expand.grid(sprintf("%04d", as.numeric(urine_random_df[,1])), c("", paste0("-",cell_placeholder[, 1])))
boxcells <- data.frame(box_cells = paste0("NEW", paste0(temp[, 1], temp[, 2])))

Which will return

   box_cells
1    NEW0010
2    NEW0011
3    NEW0012
4  NEW0010-1
5  NEW0011-1
6  NEW0012-1
7  NEW0010-2
8  NEW0011-2
9  NEW0012-2
10 NEW0010-3
11 NEW0011-3
12 NEW0012-3

If you don't like the order, you could reorder by

boxcells <- data.frame(box_cells = boxcells[order(as.numeric(substr(boxcells$box_cells, 6,7))), ])

   box_cells
1    NEW0010
2  NEW0010-1
3  NEW0010-2
4  NEW0010-3
5    NEW0011
6  NEW0011-1
7  NEW0011-2
8  NEW0011-3
9    NEW0012
10 NEW0012-1
11 NEW0012-2
12 NEW0012-3

Upvotes: 0

Related Questions