CODEWITHSUNDEEP
r
peter.petrov
peter.petrov

Reputation: 39437

R matrix - row vector - column vector - different output - confusion

Please note the following R session.

    > z3 <- cbind(c(10,10,10,10));   
    > z3
         [,1]
    [1,]   10
    [2,]   10
    [3,]   10
    [4,]   10
    > m2
         [,1] [,2] [,3]
    [1,]    1    2    3
    [2,]    4    5    6
    [3,]    7    8    9
    [4,]   10   11   12
    > m2[,2]
    [1]  2  5  8 11
    > m3
         [,1] [,2] [,3] [,4]
    [1,]   10   10   10    8
    [2,]    1    1    1    1
    > m3 %*% m2[,2];
         [,1]
    [1,]  238
    [2,]   26
    > m3 %*% z3; 
         [,1]
    [1,]  380
    [2,]   40
    >

Here m3 is a 2x4 matrix so it makes perfect sense that it can be multiplied by m2[,2] (the second column of m2) because m2[,2] is a 4x1 matrix (so they are compatible for multiplication). Now note that z3 is also a 4x1 matrix (like m2[,2]).

I would like to know why their presentations (of m2[,2] and z) are different (I mean the output I see when I just type z3 and when I just type m2[,2] at the R prompt). I find this difference strange. I like z3's presentation, I don't like m2[,2]'s presentation.

Upvotes: 1

Views: 102

Answers (1)

rbatt
rbatt

Reputation: 4807

If you subset to 1 dimension, R just turns it into a vector. (Edit: it just simplifies. I.e., "returns to lowest possible dimension")

If you don't want this to happen, do m2[,2,drop=FALSE]

As a side note, (1:4)%*%m2 works, as does m2%*%(1:3). So it's nice that even if it drops the matrix representation, it'll do the work to make your math work.

Upvotes: 3

Related Questions