PERL: Sorting Letters from A to Z - perl

Question

I'm Trying to sort letters in a file from A to Z
for example: a A B d r g
sorted: A a B d g r
#ARGV == 2 or die "Usage: $0 infile outfile\n";
open $old, '<', $ARGV[0] or die $!;
open $new, '>', $ARGV[1] or die $!;
#mass=<$old>;
#array=qw(#mass);
#sort=sort #array;
#mass1=sort {uc $a cmp uc $b} #sort;
print $new #mass1;
Where am I going wrong?

Answer 1

I don't think you understand the the standard text ordering is ASCII-based. So because all uppercase proceed all lowercase, the same is true of your input. Therefore, you order for a straight sort would be ( 'A', 'B', 'a', 'd', 'g', 'r' ).
You want to double compare the two strings. In this case, you're going to need to pass a routine to sort.
#sort= sort { lc $a cmp lc $b or $a cmp $b } #array;

Answer 2

I'm not sure what you intended to do with qw, but
suffice it to say that the contents of #mass will be never be used.
#array = qw(hello world);
Will cause #array to be defined to contain 2 strings, hello and world. It is just shorthand for:
#array = ('hello', 'world');
Which is why
#array=qw(#mass);
Evaluates to ('#mass') - an array with the single literal string of 5 characters #mass.
Maybe that's what you're doing wrong. What if you try
#array = map { split /\s+/} #mass;
#mass is the list of lines. Each line has words or just letters, separated by space.
What that last line does is maps each line with split /\s+/ - which will split each
line like 'ba ab a G' into a list like ('ba', 'ab', 'a', 'G') and #array will
become a single list of words/letters.
Then it's a matter of how you want to sort them. See the other answer as well.
Oh, and remember to put back the spaces when you write out your file:
print $new (join " ", #mass1);
If you want each line to be sorted interdependently of the other, that's easy too:
$mass1 = join "\n", map { join " ", sort (split /\s+/) } #mass
That reads, 'for every line in #mass, split on space, sort and join back again with space', and with the resulting array, join with newline to produce the output of the file.
Note that you can drop in sort with a comparator like sort { $a cmp $b } etc.
If your file is too big, then looping is maybe prudent:
for my $mass (<$old>) {
my $sorted_line = join " ", sort (split /\s+/, $mass);
print $new "$sorted_line\n";
}

Answer 3

You need to find the correct LOCALE to use, so that the order used by all functions (sort, etc) are using the correct locale and sort accordingly to it.
See this page showing most of the variables defining locales, and look for LANG and LC_ALL. and LC_COLLATE (I have to admit I'm not exactly sure which is used when. LC_ALL is supposed to take precedence over the others, so it's the one you can change to have all LC_* values set... Please test, ymmv)
I believe you probably need to use one of the unicode locales. Ascii won't do what you want, as CAPS are before regular letters in ascii.
To find out which locales you can use: locale -a
To see which locales you are currently set to : locale (user and system-wide values are possible)
You probably need something containing "utf-8" to have the order you seek
Then : (if for example en_US.UTF-8 is available):
just before using it in the sort, define locales you want to sort with:
LC_ALL=en_US.UTF-8
(or whatever the value you need it to be set at, and is available as shown by "locale -a")
(save/restore their previous values around the invocation if you need to)
In shell, you probably better want to ass "export" to those variables you redefine, to ensure subshells use the new value too (like: something | sort : in bash, sort will be in a subshell, therefore using the default value of LC_*, or using the exported value if you exported it!)