CODEWITHSUNDEEP
rforeachdoparallel
giupardeb
giupardeb

Reputation: 841

error calling combine function loop foreach in R

The following code:

df <- foreach(i = 1:length, .combine = cbind) %dopar% {
 ...
 ...
 m4
}

gives me an error :

error calling combine function:
<simpleError in data.frame(..., check.names = FALSE): arguments imply differing number of rows: 17, 0>

but, if I execute the same code with "for" instead of "foreach" and I also make the cbind manually:

for (i in 1:lengthGeni) {
...
...
mFinaleCGunificati <- cbind(mFinaleCGunificati, m4)
}

everthing works well

Upvotes: 3

Views: 1728

Answers (1)

loki
loki

Reputation: 10340

From ?cbind we learn that

If there are several matrix arguments, they must all have the same number of columns (or rows) and this will be the number of columns (or rows) of the result. If all the arguments are vectors, the number of columns (rows) in the result is equal to the length of the longest vector.

However if from the comparison of these two cbind examples

# second df with one row
cbind(as.data.frame(matrix(rep(1, 10), nrow = 5)), 
      as.data.frame(matrix(rep(2, 2), nrow = 1)))
#   V1 V2 V1 V2
# 1  1  1  2  2
# 2  1  1  2  2
# 3  1  1  2  2
# 4  1  1  2  2
# 5  1  1  2  2

# second df with zero rows
cbind(as.data.frame(matrix(rep(1, 10), nrow = 5)), 
      as.data.frame(matrix(rep(2, 0), nrow = 0)))
# Error in data.frame(..., check.names = FALSE) : 
#   arguments imply differing number of rows: 5, 0

we learn that zero-length objects are not allowed.

Therefore, you should check if your result in the loop has any number of rows greater than 0. 

library(foreach)
library(doSNOW)
cl <- makeSOCKcluster(5)
registerDoSNOW(cl)

df <- foreach(i = 1:2, .combine = cbind) %dopar% {
  if (i == 1){
    x <- as.data.frame(matrix(rep(1, 5), nrow = 5))
  } else {
    x <- as.data.frame(matrix(rep(1, 2), nrow = 1))
  }

  # check if result has at least one row
  if (nrow(x) > 0){
    x
  }
}
df 
#   V1 V1 V2
# 1  1  2  2
# 2  1  2  2
# 3  1  2  2
# 4  1  2  2
# 5  1  2  2

However, keep in mind that shorter vectors will be reused. Thus this approach might lead to redundancies within the code.

To aviod redundancies, you might consider matching the lengths of the results before returning it in the foreach loop.

Upvotes: 1

Related Questions