foreach my $key ( grep { /^[A-Z]+$/ } keys %runset ) {
    print $key, " ", $runset{$key}, "\n";
}
[download]

... But if you really wanted to delete the non-uppercase hash elements, then expanding on this usage of grep ...

delete $runset{$_} for ( grep { !/^[A-Z]+$/ } keys %runset );
[download]

perl -e 's&&rob@cowsnet.com.au&&&split/[@.]/&&s&.com.&_&&&print'

Minor regex nit:

Rather than checking the entire key for having uppercase letters and then negating the rest, you could check for the presence of any non-uppercase letter:

delete $runset{$_} for ( grep { /[^A-Z]/ } keys %runset );

Of course, this will also allow the null hash key as well, so you might want to extend the grep to grep { $_ && /[^A-Z]/ }, and by that point, you may have sacrificed the savings on the regex.

- Richie

for (keys %runset) {
  unless($_ eq uc($_)) { 
    delete $runset{$_};
  } 
}
[download]

However you have another problem: a hash is un-ordered, so you won't get the sentence you expect unless you are very lucky - you need some way to specify the order in which the values of the uppercase keys are printed. The example you provide doesn't work for a simple sort of the uppercase keys so you'll have to think of something else.

--
I'd like to be able to assign to an luser

While it goes beyond the scope of the problem, a hash can be made to be ordered by using the CPAN module Tie::IxHash.

Link to Tie::IxHash documentation
Link to Tie::IxHash tarball


Usage:

#!/usr/bin/perl -w

use strict;
use Tie::IxHash;

tie my %ordered_hash, Tie::IxHash;

%ordered_hash = (a => 1, b => 2, c => 3);

foreach my $key ( keys %ordered_hash ) {
    print "$key\t$ordered_hash{$key}\n";
}

# prints:
# a     1
# b     2
# c     3

exit 0;

Cheers,
Richard

Actually, as long as your print statement includes $key, your output is also not going to be formatted quite as you expect: "TEST IABC amBAR a monk in training" is one possibility. (Note the lack of spaces, too.) What you really want inside the loop is probably something like this:
print $runset{$key}, " "

Or, even more Perlish, get rid of the enclosing loop and just use this:
print join " ", values %runset;

This latter version also eliminates the trailing space, because join only inserts delimiters between list elements.

This is actually much easier than you might expect - Essentially you are looking at merging two hashes with the key-values of %runset having precedence of those in %description. This can be done as follows:

delete $runset{$_} for ( grep { !/^[A-Z]+$/ } keys %runset );

%merged_hash = ( %description, %runset );
[download]

This code will delete the non-uppercase key-value pairs in the %runset and then merge the two hashes - Any values in %description with an identical key-name in %runset will be replaced with the values in the latter (based on left-to-right execution).

 

perl -e 's&&rob@cowsnet.com.au&&&split/[@.]/&&s&.com.&_&&&print'

Hi Rob,

Thanks for sharing your knowledge. Even in your .sig?

I tried pasting your .sig:

$
$ perl -e 's&&rob@cowsnet.com.au&&&split/[@.]/&&s&.com.&_&&&print'
In string, @cowsnet now must be written as \@cowsnet at -e line 1, nea
+r "rob@cowsnet"
Execution of -e aborted due to compilation errors.
$ perl -e 's&&rob\@cowsnet.com.au&&&split/[@.]/&&s&.com.&_&&&print'
rob@cowsnet_au$
[download]

Am I missing something?

HTH, Danny

Update:

$ perl -v
 
This is perl, v5.6.0 built for i586-linux
 
Copyright 1987-2000, Larry Wall
.
.
.
[download]

Not sure if this is the *best* way, but I'm lazy and it works for me

#!/usr/bin/perl

use strict;

my %h = qw/A a B b c c/;
my %h2;

$h2{$_} = $h{$_} for grep {/^[^a-z]$/} keys %h;
[download]