R help - Dec 1999 - R as Unix-Filter and Streams for DataMining

Hi,
i want to use R as a filter in a Unix-like way. So i want to start it
with

producerTool | R --no-save --no-restore --slave < Rscript |
anotherConsumerUnixTool

inside the Script i put the results to stdout, works fine, but how to
read in  
from stdin? Its obvious that i have to source the Rscript in a different
way. Any Hints?

My second question is, if there is a possibility to use something like
the Scheme Streams
(see Abelson/Sussman "Structure and Interpretation of Computer
Programs") for input of Huge Amounts of Data, that do not fit in the
main memory, or to reduce the necessary amount of memory? The
application
i have in mind is: Generate from a training set a linear or other model
and then 
pipe in a file with the test cases (as stream with read.table) and give
the output to stdout. Is there 
any concept in R like this?

Thanks in advance
  Chris 
-- 
Christoph M. Friedrich        | mailto:friedrich at computer.org 
Gesellschaft f?r Modulfermenterbau mbH (GfM mbH) |
http://www.tussy.uni-wh.de/~chris
Alfred-Herrhausen Str. 44 ; D-58455 Witten, Germany
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

My workaround currently is:

(1) Write the path and name of your input file into a file with a *fixed*
name
    use unix paths //// never windows \\\\, e.g.

  d:/temp/filename.txt

(2) Call R via Batchfile

(3) In the batch file set a temporary user dir for R

  set R_USER=d:/temp

(4) Have an approbriate .Profile in this temporary user dir, which defines
.First
    Do *not* take other actions in .Profile except for setting options and
defining .First

.First <- function(){
  # get filename
  filename <- scan("d:/temp/filename.txt", what=character(0))
  cat("filename is ", filename, "\n")
  # make use of filename
  source(filename)
  # exit from R
  q()
}


If you discover a simpler solution, please give me a note
Best


-----Original Message-----
From: owner-r-help at stat.math.ethz.ch
[mailto:owner-r-help at stat.math.ethz.ch]On Behalf Of Christoph M.
Friedrich
Sent: Friday, December 03, 1999 1:38 PM
To: r-help at stat.math.ethz.ch
Subject: [R] R as Unix-Filter and Streams for DataMining


Hi,
i want to use R as a filter in a Unix-like way. So i want to start it
with

producerTool | R --no-save --no-restore --slave < Rscript |
anotherConsumerUnixTool

inside the Script i put the results to stdout, works fine, but how to
read in
from stdin? Its obvious that i have to source the Rscript in a different
way. Any Hints?

My second question is, if there is a possibility to use something like
the Scheme Streams
(see Abelson/Sussman "Structure and Interpretation of Computer
Programs") for input of Huge Amounts of Data, that do not fit in the
main memory, or to reduce the necessary amount of memory? The
application
i have in mind is: Generate from a training set a linear or other model
and then
pipe in a file with the test cases (as stream with read.table) and give
the output to stdout. Is there
any concept in R like this?

Thanks in advance
  Chris
--
Christoph M. Friedrich        | mailto:friedrich at computer.org
Gesellschaft f?r Modulfermenterbau mbH (GfM mbH) |
http://www.tussy.uni-wh.de/~chris
Alfred-Herrhausen Str. 44 ; D-58455 Witten, Germany
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.
-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._.
_._

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

"Christoph M. Friedrich" wrote:
> Hi,
> i want to use R as a filter in a Unix-like way. So i want to start it
> with
>
> producerTool | R --no-save --no-restore --slave < Rscript |
> anotherConsumerUnixTool
>
> inside the Script i put the results to stdout, works fine, but how to
> read in
> from stdin? Its obvious that i have to source the Rscript in a different
> way. Any Hints?
Two (klugey) ideas--

- write a script which reads standard input, and reads a file, and outputs
both of them to standard output.  (Maybe there's a command that does this
already; anyone know?)  You could take this one step further and just make
it a wrapper for R, which would execute any files given on the command
line and then read from standard input.

- put your script (or a command to load your script) into .Rprofile, which
R reads and executes initially.

Sorry, I don't know anything about streams.

-John

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._