Strange (non-deterministic) problem with strsplit

"HenrikB" == Henrik Bengtsson <hb@maths.lth.se>
    on Sat, 17 Jul 2004 01:59:17 +0200 writes:

gsub(" ",   "", "b c + d | a * b", perl=TRUE)

gsub("\\s", "", "b c + d | a * b", perl=TRUE)

gsub("\\s", "", "bc + d | a * b", perl=TRUE)

gsub(" ",   "", "bc + d | a * b", perl=TRUE)

BTW,
what's the name of the error logger for Windows that is once in a while
recommended on this list and that gives more detailed errors than the
default Windows one?)

[Moving this thread to R-devel instead]

I suspect your "random" results are due to a bug in gsub(). On my R v1.9.0
(Rterm and Rgui) R crashes when I do

% R --vanilla

gsub(" ", "", "abb + c | a*b", perl=TRUE)

Trying 

gsub(" ", "", "b c + d | a * b", perl=TRUE)

and I'll get NULL. With 

gsub("\\s", "", "bc + d | a * b", perl=TRUE)

it works as expected. So there is something buggy for sure.

This might have been fixed in R v1.9.1 or its patched version. (I'm still
busy to recover from a HDD crash, but, yes, I will update to Rv1.9.1. BTW,
what's the name of the error logger for Windows that is once in a while
recommended on this list and that gives more detailed errors than the
default Windows one?)

Cheers

Henrik

-----Original Message-----
From: r-help-bounces@stat.math.ethz.ch 
[mailto:r-help-bounces@stat.math.ethz.ch] On Behalf Of hadley wickham
Sent: Friday, July 16, 2004 10:50 PM
To: r-help@stat.math.ethz.ch
Subject: [R] Strange (non-deterministic) problem with strsplit


I'm having an odd problem with strsplit (well I think it's 
strplit that's causing the problem).  When I run the code 
below as follows:  str(parseFormulaMin(y +x +d ~ b +d +e| a * b))

I expect to get
List of 3
 $ y: chr "y+x+d"
 $ x: chr "b+d+e"
 $ g: chr "a*b"

But about half the time I get 

List of 3
 $ y: chr "y+x+d"
 $ x: chr "b+d+e"
 $ g: chr "a*[square box]"
(square box not reproduced here because copy and pasting it 
seems to break my web mail)

Can anyone reproduce the problem and/or suggest any solutions? 

parseFormula <- function(formula) {
	splitvars <- function(x) {
		strsplit(x, "\\+|\\*")[[1]]
	}
	stripwhitespace <- function(x) {
		gsub("\\s", "", x, perl=T)
	}
	
	vars <- stripwhitespace(as.character(formula)[3])
	varsplit <- strsplit(vars, "|", fixed=TRUE)[[1]]

	parts <- list(
		y = stripwhitespace(as.character(formula)[2]),
		x = varsplit[1],
		g = varsplit[2]
	)
	lapply(parts, splitvars)
} 

Thanks,

Hadley