Skip to content
Prev 174119 / 398503 Next

the quote problem with readLines()

The amount of data that you want to read in (136M numbers) will
require about 1GB of memory (8 bytes per number for floating point -
truncation does not reduce this number of bytes).  So if you want to
read it all in, then find a 64-bit version of R and probably at least
4GB of memory for your process.  A 32-bit version might have just
enough space if you can allocate all the 4GB of memory to that
process.

So if you want to have it all in memory, invest in a larger computer.
If you want to run on the system you have, then you will probably have
to sample your data so that you can get a portion that will fit in
memory to run your test, or see if there is a way of processing
portions of the file and then combining for a final result.
On Wed, Mar 18, 2009 at 9:58 AM, Dongyan Song <yzhskdls at hotmail.com> wrote: