CODEWITHSUNDEEP
rmatrixsparse-matrix
Stephen
Stephen

Reputation: 2424

R coordinate-value pairs into a sparse matrix

I'm having trouble loading my data set into a sparse matrix in R. I am using the Matrix package. The data I have is in the form x y value. For example:

  V1 V2 V3
  1  2  .34
  7  4  .56
  4  5  .62

where I would want to do the equivalent of

myMatrix[1,2] = .34
myMatrix[7,4] = .56 
myMatrix[4,5] = .62

in an automated fashion.

I want to do something like:

myMatrix = Matrix(nrow=numrows, ncol=numcols)
myMatrix[mydata[1:numrows, 1], mydata[1:numrows, 2]] <- mydata[1:numrows, 3]

but this makes my matrix a lgeMatrix when I need a numeric matrix.

I have also tried:

myMatrix = Matrix(nrow=numrows, ncol=numcols)
for(i in 1:numrows){
    myMatrix[mydata[i, 1], mydata[i, 2]] <- mydata[i, 3]
}

Which creates the kind of matrix I want, but it takes way too long (more than 5 minutes). I know it works because when I stop it I check the first few values and they're correct, but the last values are NA. I'm working with a 7095 by 5896 matrix with 247158 values to enter so a for loop is out of the question, unless I'm just being impatient.

My question is: What is the preferred way to do this in R?

update:

I figured it out using sparseMatrix instead:

myMatrix = sparseMatrix(i = mydata[1:numrows,1], j = mydata[1:numrows,2],
                     x = mydata[1:numrows,3])

didn't understand the sparseMatrix usage in the other post

Upvotes: 3

Views: 1517

Answers (1)

IRTFM
IRTFM

Reputation: 263332

Let's assume that is a dataframe named dat:

myMatrix = Matrix(0, nrow=10, ncol=10)
# Notice that you need to specify zero values to make it sparse.
myMatrix[cbind(dat$V1, dat$V2)] <- dat$V3
myMatrix
#--------------
10 x 10 sparse Matrix of class "dgCMatrix"

 [1,] . 0.34 . .    .    . . . . .
 [2,] . .    . .    .    . . . . .
 [3,] . .    . .    .    . . . . .
 [4,] . .    . .    0.62 . . . . .
 [5,] . .    . .    .    . . . . .
 [6,] . .    . .    .    . . . . .
 [7,] . .    . 0.56 .    . . . . .
 [8,] . .    . .    .    . . . . .
 [9,] . .    . .    .    . . . . .
[10,] . .    . .    .    . . . . .

Upvotes: 6

Related Questions