If you want something that perhaps is more readable but probably less
efficient & certainly less general than Derek's solution this one might
work:
use PDL::NiceSlice;
sub setme_jb {
my ($x, $y, $value) = @_;
# Error checking non-existent.
# Find those positions that are != 0
my $ii = which($found->index2d($x, $y) != 0);
if ($ii->nelem() == 0) {
return;
}
# Set the relevant parts.
$found->index2d($x($ii), $y($ii)) .= $value;
}
This is really written as a 2D version and doesn't generalise very
well, but I thought you might find it useful as an alternative approach
(extending this using whichND and indexND is a non-trivial task btw as
it seems like whichND is a bit broken...)
Cheers,
J.
On 3 Jan 2008, at 23:30, [EMAIL PROTECTED] wrote:
Quoting Derek Lamb <[EMAIL PROTECTED]>:
And since indexND is implemented as a call to range, this might work
for
you.
It does! Thanks. My real data has sped up from about 500 loops per
hour to
8000 in 4 minutes. Obviously a massive improvement.
It may or may not be easier for you to read at the moment!
Right now, it's totally incomprehensible. But I'm no longer kicking a
dead
Lamborghini down the highway, and learning from code which works is
always
easier than trying to code something you don't understand.
_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
