Performing a non parametric Friedman Test

My data looks like this:
                              (treatments)
                 T1             T2            T3
DK              8               5               3 

JP              5                4              1

AS             9                7             4

MK             8                4              4

DK, JP, AS, and MK are 4 different people (blocks) I am using.

This is my code so far:

treat<-c("J","J","J","J","U","U","U","U","H","H", "H", "H")

y<-c(8,5,3,5,4,1,9,7,4,8,4,4)

block<-c("DK","JP","AS","MK","DK","JP","AS","MK","DK","JP","AS","MK")

ranks<-matrix(NA,nrow=4,ncol=3)

for (i in 1:4) ranks[i,]<-??????????????????

I know that once I can get the ranks stored I can do a command like below.

summary(aov(as.vector(ranks)~factor(treat)+factor(block)))

Can someone help me store the data as ranks?