※ 引述《untitled (Causality)》之铭言：
: ※ 引述《mapleV (Sucks)》之铭言：
: : I found a new way to transform usps and it only took 3x seconds.
: !!! How's the CPU of your computer?
PIII 733Mhz for executing the transformation from loaded data to
matrix. 3x seconds.
Loading data takes 8~10 seconds.
: Would you like to share your code with us next Monday?
I think it's not necessary to show my code next Monday.
My code is customized for usps data. Not very general.
Just a simple idea.
If there exists any errors, please remark it.
load_transform <- function(filename , ncolumn)
{
mdt <- readLines(filename)
temp <- list()
for (i in 1: length(mdt))
temp[[i]] <- as.numeric(matrix(strsplit(mdt[i], " |:")[[1]][2*c(1:ncolumn)-1]))
dx <- matrix(,ncol = ncolumn ,nrow = length(mdt))
for (i in 1:length(mdt))
dx[i,] <- temp[[i]]
dx
}