[R] naive question

Tony Plate tplate at blackmesacapital.com
Wed Jun 30 19:23:37 CEST 2004

To be careful, there's lots more to I/O than the functions read.table() & 
scan() -- I was only commenting on those, and no inference should be made 
about other aspects of S-plus I/O based on those comments!

I suspect that what has happened is that memory, CPU speed, and I/O speed 
have evolved at different rates, so what used to be acceptable code in 
read.table() (in both R and S-plus) is now showing its limitations and has 
reached the point where it can take half an hour to read in, on a 
readily-available computer, the largest data table that can be comfortably 
handled.  I'm speculating, but 10 years ago,  on a readily available 
computer, did it take half an hour to read in the largest data table that 
could be comfortably handled in S-plus or R?  People who encounter this now 
are surprised and disappointed, and IMHO, somewhat justifiably so.  The 
fact that R is an open source volunteer project suggests that the time is 
ripe for one of those disappointed people to fix the matter and contribute 
the function read.table.fast()!

-- Tony Plate

At Wednesday 10:08 AM 6/30/2004, Igor Rivin wrote:

>Thank you! It's interesting about S-Plus, since they apparently try to support
>work with much larger data sets by writing everything out to disk (thus 
>around the, eg, address space limitations, I guess), so it is a little 
>that they did not tweak the I/O more...
>         Thanks again,
>                 Igor

More information about the R-help mailing list