Multiply matrix with a vector row wise

Question

I've got a logical matrix and I need to multiply each column by the sum of this column using apply. For example :

> a
     [,1] [,2] [,3] [,4]
[1,]    1    1    1    1
[2,]    0    0    0    0
[3,]    1    1    0    1
[4,]    1    0    0    1
> b <- colSums(a)
> b
[1] 3 2 1 3

And I want to get the following matrix :

 > a
     [,1] [,2] [,3] [,4]
[1,]    3    2    1    3
[2,]    0    0    0    0
[3,]    3    2    0    3
[4,]    3    0    0    3

I did it with for but since I need to apply my function to a huge dataset I need to code with apply. Thank you.

It does but it takes a long time to run. If I use the function apply of biganalystics it won't take long time, that's why I needed to use apply. — Taoufiq Mouhcine, Aug 03 '15 at 13:48

akrun · Answer 1 · 2015-08-03T13:34:25.687

3

You can take the transpose (t) of the matrix 'a' and then multiply with the vector ('b'), take the transpose (t) of the output to get the desired result.

 t(t(a)*b)

Or we can make the lengths of the 'a' and 'b' same by replicating the elements of 'b'. By doing b[col(a)], we get each element of 'b' replicated by the index provided by the col.

 a*b[col(a)]

For better understanding

  col(a)
  #     [,1] [,2] [,3] [,4]
  #[1,]    1    2    3    4
  #[2,]    1    2    3    4
  #[3,]    1    2    3    4
  #[4,]    1    2    3    4

 b[col(a)] #is a vector
 #[1] 3 3 3 3 2 2 2 2 1 1 1 1 3 3 3 3

 a*b[col(a)]
 #     [,1] [,2] [,3] [,4]
 #[1,]    3    2    1    3
 #[2,]    0    0    0    0
 #[3,]    3    2    0    3
 #[4,]    3    0    0    3

edited Aug 03 '15 at 13:34

answered Aug 03 '15 at 13:27

akrun

874,273
37
540
662

1

Thanks for your answer akrun. I tried that but R cannot run that operation because I have a huge matrix (bigmemory) – Taoufiq Mouhcine Aug 03 '15 at 13:45
@TaoufiqMouhcine In that case, the `apply` would be more appropriate as the big matrix will take a toll on the memory. I didn't read the part where you mentioned big data (sorry!) – akrun Aug 03 '15 at 13:47

score 1 · Accepted Answer · answered Aug 03 '15 at 13:39

In addition to @akrun's answer, if you really do want to use apply:

apply(a,2,function(x)x*sum(x))
#     [,1] [,2] [,3] [,4]
#[1,]    3    2    1    3
#[2,]    0    0    0    0
#[3,]    3    2    0    3
#[4,]    3    0    0    3

2 means you work on columns (i. e. the second dimension). So each operation is done on a vector corresponding to a column, hence the use of sum (which works on vector) instead of colSums (which works on a matrix).

Multiply matrix with a vector row wise

2 Answers2