Search code examples
rarrayssparse-matrixmemory-efficient

How to add elements within an array based on information from another array in R?

I have an array in which I need to add elements together based on the index in which they belong. 

row_index <- c(1, 3, 1, 2, 2, 3, 3, 4)
array <- c(1.2, 3.4, 5.6, 7.2, 0.3, 5.6, 2.0, 3.1)

So in this case I want to add all elements that belong on the first row according to row_index together. For example, 1.2 and 5.6 would be added together since they correspond to the same row based on row_indexand similarly for row 2 it would be 7.2 + 0.3. So my resulting array would be.

result = (6.8 , 7.5 , 11.0 , 3.1)

I have no clue how to do this in R. For context, I have a sparse matrix and accessing which row each non-zero elements belongs to is very easy and efficient to do, and the length of row_index will always be the same as array.

Solution

  • One option is tapply from base R where the grouping is based on 'row_index' 

    unname(tapply(array, row_index, FUN = sum))
#[1]  6.8  7.5 11.0  3.1

    Or another is by

    by(array, row_index, FUN = sum)

    Or with rowsum

    unname(rowsum(array, row_index)[,1])
#[1]  6.8  7.5 11.0  3.1

    Or with xtabs

    xtabs(array ~ row_index)

    NOTE: All of the above solutions are from base R