Make weighted edge list like pairs to data frame in R

Question

I have a data set with three columns.

X = c("x", "y", "z", "x", "y", "w")
Y = c("a", "b", "c" , "b", "a", "c")
set.seed(3)
Z = sample(10,6)
dat = data.frame(X, Y, Z)

The dat looks like

 X Y Z
1 x a 2
2 y b 8
3 z c 4
4 x b 3
5 y a 9
6 w c 6

I want to transform the above data into the format that the rows are X and the columns are Y and the values in the matrix are Z. Essentially I want the following:

    a   b    c
x   2   3    0
y   9   8    0
z   0   0    4
w   0   0    6

This is somewhat similar to edge list to adjacency matrix, except the data matrix is not square. Please help. Thanks!

Answer 1

Try

library(reshape2)
acast(dat, X~Y, value.var='Z', fill=0)

To get the rows in order as the expected result

acast(dat, factor(X, levels=unique(X))~Y, value.var='Z', fill=0)

Or as @thelatemail commented

xtabs(Z~X+Y, data= dat)[unique(dat$X),]

Or a modified version if both columns are not in order

with(dat, xtabs(Z ~ X + Y)[unique(X), unique(Y)])