The problem is I don't know how to spawn the instance of S-Plus then at a later
point ask it to execute a script. Right now I'm just spawning the process with
the /BATCH flag and passing it a script. It runs, exits then I scan a directory
for results.
Thanks
-----Original Message-----
From: A.J. Rossini [mailto:blindglobe@gmail.com]
Sent: Thursday, April 13, 2006 10:42 AM
To: Brian Richards
Cc: s-news@lists.biostat.wustl.edu
Subject: Re: [S] Executing S-Plus scripts
Why not just use a "thread pool" of sorts to keep a few running "in
preparation"? I've seen this done in the past with R, and it should
be possible (modulo licenses) to do with S-PLUS.
Both Java and Python provide some assistance with this approach, I'm
sure .Net would as well.
On 4/13/06, Brian Richards <brichards@pharsight.com> wrote:
>
> I'm investigating running S-PLUS scripts through a seperate application.
> Currently we are using command line invocation of splus.exe to accomplish
> this. But there seems to be alot of overhead with starting the process
> running it and producing results. I was wondering if anyone had any other
> ideas on how I could speed the execution time up. We don't need direct
> access to splus functions via the application. Simply want to run a script
> and read output from disk. We're using .Net. I looked into sconnect but it
> seems to maybe be a deeper level of integration than we require at this
> time.
>
> Optionally, if there is a way to start the engine and keep it running and
> then run scripts contiguously then shutdown the engine afterwards that would
> probably work too. In the future we'd like to be able to generate and read
> dataframes, within out .Net framework (managed code not just using the
> newest compiler to generate unmanaged code). I'm very green with S-Plus, and
> have looked through the archives but I could only find posts for compiling
> sconnect with VS .Net.
>
> Thanks
>
>
> Brian Richards
--
best,
-tony
blindglobe@gmail.com
Muttenz, Switzerland.
"Commit early,commit often, and commit in a repository from which we can easily
roll-back your mistakes" (AJR, 4Jan05).
|