I have a 32000 x 14 matrix (M) where entry (x, y) corresponds to
person x at a discrete time y. I then have a matrix (M2) of 60000 x 2
where each entry is a an event by a person that is in the above 32000
and at a time that is in the range of the discrete time points above.
I want to populate the another matrix (M3) such that (x, y) is the
number of events of person x between times y-1 and y. This is simple
to code, (I just put the discrete time points and the 32000 people in
their own vectors, V1 and V2):
for (i in 1:32000){
for (j in 1:60000){
for (k in 2:14){
if (M2[1, j] == V2[i] & M2[2, j] > V1[k-1] & M2[2, j] <=
V1[k])
M3[i, k] = M3[i, k]+1
}}}
This would work im sure, the problem is that it takes a very long
time. Any advice on helping it run faster would be greatly
appreciated.
without a reproducible sample, it is hard to tell, but I will give it a shot. Maybe it's possible to merge your M with M2: merge(M, M2) If you only want to count times, you can use seq_along(x) in a by function eg: dat <- data.frame(person=rep(c(1,2,3), each=5), time=rnorm(15)) by(dat$person, dat$person, function(x) seq_along(x)) Ofcourse you can use any of the other vectorisation functions (lapply, apply, aggregate, ....) HTH Bart -- View this message in context: http://r.789695.n4.nabble.com/Counter-in-a-For-Loop-Efficiency-Issue-tp3253674p3253698.html Sent from the R help mailing list archive at Nabble.com.
On Wed, Feb 02, 2011 at 09:35:28AM +0200, Leendert Punt wrote:> I have a 32000 x 14 matrix (M) where entry (x, y) corresponds to > person x at a discrete time y. I then have a matrix (M2) of 60000 x 2 > where each entry is a an event by a person that is in the above 32000 > and at a time that is in the range of the discrete time points above. > I want to populate the another matrix (M3) such that (x, y) is the > number of events of person x between times y-1 and y. This is simple > to code, (I just put the discrete time points and the 32000 people in > their own vectors, V1 and V2): > > > for (i in 1:32000){ > for (j in 1:60000){ > > for (k in 2:14){ > > if (M2[1, j] == V2[i] & M2[2, j] > V1[k-1] & M2[2, j] <= V1[k]) > M3[i, k] = M3[i, k]+1 > > }}}Consider the following approach, which uses M2, V1 and V2 as used in the code above. V1 <- c(0, 10, 20, 30, 40) V2 <- 1:3 # list of persons in the required order M2 <- rbind( c(1, 2, 1, 1, 3, 1, 3, 1), c(11, 21, 21, 21, 11, 21, 11, 31)) person <- factor(M2[1, ], levels=V2) interval <- cut(M2[2, ], breaks=V1) M3 <- table(person, interval) M2 [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [1,] 1 2 1 1 3 1 3 1 [2,] 11 21 21 21 11 21 11 31 M3 interval person (0,10] (10,20] (20,30] (30,40] 1 0 1 3 1 2 0 0 1 0 3 0 2 0 0 Does this approach work for your data? Petr Savicky.