On 02/07/09 21:02, Mark Wardle wrote:
Hi.
I've stared at your code for a number of minutes and can't understand
what you're trying to achieve here. It looks as if you're fighting
scope - it may be worth refactoring your approach to simplify matters.
Otherwise, it sounds like a recipe for obfuscation! What are you
trying to do really?
b/w
Mark
2009/7/2 Allan Engelhardt<all...@cybaea.com>:
Must be the heat or something but I can't get my brain into gear and figure
out how to get something like
if (1) { c<- 1; foo<- function () print(c); }
Imagine "c <- 1" is a long complex operation, say to get a list of
current active EC2 hosts through querying some web services and some
complex system("ssh...") calls that I do not want to repeat ever again,
and the result is a list of hosts (and it is not really called "c" -
that would be bad). My function foo is a parallel version of lapply and
the code goes something like
if ( usingAmazonWebServices() ) {
cluster <- myServers(); # Takes minutes
p.lapply <- function (...) aws.lapply(..., hosts = cluster)
} else if ( require("multicore", ...) ) {
p.lapply <- function (...) mc.lapply(..., mc.preschedule = FALSE)
} else {
p.lapply <- lapply
}
## ...
## Much later:
result <- p.lapply(problems, analysis)
This is somewhere far away from the main action of the code and my fear
is that somewhere someone (most likely me next week) will accidentally
use the "custer" variable for something else, causing p.lapply to
mysteriously fail. I probably should package it all up and use
namespaces or some such, but I am just trying not to get hurt right now.
Allan
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
