Faster Printing Alternatives to 'cat'

Do it in chunks of rows. /H

On Sat, Jan 17, 2009 at 6:55 AM, Gundala Viswanath <gundalav at gmail.com> wrote:

Hi Hadley,

I had to do it by line. Because, in practice, I will
manipulate the figures and string before printing it.

And I can't bind these results into one new object,
because there are literally millions of this lines, and
R object can't handle that in my 4GB Ram memory.

I tried your suggestion already, I have memory problem

x <- cbind(dat$V1, as.character(dat$V2))

Error: cannot allocate vector of size 4.2 Gb
Execution halted

- Gundala Viswanath
Jakarta - Indonesia



On Sat, Jan 17, 2009 at 11:39 PM, hadley wickham <h.wickham at gmail.com> wrote:

On Sat, Jan 17, 2009 at 7:59 AM, gundalav <gundalav at gmail.com> wrote:

Dear Jim and all,

Allow me to ask your expert opinion.


Using the data (16Mb) downloadable from here:

http://drop.io/gundalav/asset/test-data-zip


It took this long under 1994.070Mhz Cpu Linux, using
"write.table"

proc.time() - ptm1

    user    system   elapsed
16581.833  5787.228 21386.064



__MYCODE__

args <- commandArgs(trailingOnly=FALSE)
fname <- args[3]
dat <- read.delim(fname, header=FALSE);

output <- file('output_writetable.txt', 'w')


ptm1 <- proc.time()
for (i in 1:nrow(dat)) {

    #cat(dat$V1[i]," ", as.character(dat$V2[i]),"\n", sep="")
    write.table(cbind(dat$V1[i], as.character(dat$V2[i])),
file=output, sep="\t", quote=FALSE, col.names=FALSE, row.names=FALSE)
}

close(output)
proc.time() - ptm1
__END__

Perhaps I misunderstood you. But seems that this is
truly slow. Is there a way I can speed it up?

Don't do it line by line!

write.table(dat[, c("V1", "V2")], file='output_writetable.txt',
sep="\t", quote=FALSE, col.names=FALSE, row.names=FALSE)

Hadley

--
http://had.co.nz/