Numerically sort a file on a given column where column is a $var - Perl

On Thu, 17 Jul 2008 14:29:52 -0700 (PDT), joemacbusiness@yahoo.com wrote:

>Hi All,
>
>I want a subroutine that will sort a file on any given column.

Maybe your just in a rush, but I stopped reading here, when I read "sort a file"

sln

>>>>> "j" == joemacbusiness <joemacbusiness@yahoo.com> writes:

j> The problem is that I cannot get the bynum() to accept 2 args.
j> The code will work if I hard-code the $col in bynum, and
j> tweek the "problem line" a bit but that defeats the purpose of the
j> "sort on column" feature.

j> open(F,"$infile") || print "cannot open $infile $!";

useless use of quotes on a scalar. not needed and it can be a bug.

j> my @array = <F>;
j> close(F);

use File::Slurp ;
my @data = read_file( $infile ) ;

j> numericSortCol(1,\@array);

j> sub numericSortCol {
j> my $col = shift;
j> my $aref = shift;
j> foreach my $item (sort bynum($col,@{ $aref })){ # <<<< problem
j> line??

well, that isn't how perl's sort works. it can take a sub name (you have
a sub call) or a code block.


and you should look at Sort::Maker which can do this for you
easily. just create a simple expression to get the column you want based
on $_. that could be something like:

(split( ' ', $_))[$col]

the rest i leave as an exercise.

uri

--
Uri Guttman ------ uri@stemsystems.com -------- http://www.sysarch.com --
----- Perl Code Review , Architecture, Development, Training, Support ------
--------- Free Perl Training --- http://perlhunter.com/college.html ---------
--------- Gourmet Hot Cocoa Mix ---- http://bestfriendscocoa.com ---------

joemacbusiness@yahoo.com wrote:
> Hi All,

> I want a subroutine that will sort a file on any given column.
> So I have a file like this:

> 300 400 500 600 700
> 33 2337483 482 78374 4567
> 10 20 30 40 50
> 1000 1001 1002 1003 1004
> 9 8 7 6 5

> And I run my numericSortCol() routine on it sorting on
> column 1 and should get this:

> 9 8 7 6 5
> 10 20 30 40 50
> 300 400 500 600 700
> 1000 1001 1002 1003 1004
> 33 2337483 482 78374 4567

> The problem is that I cannot get the bynum() to accept 2 args.
> The code will work if I hard-code the $col in bynum, and
> tweek the "problem line" a bit but that defeats the purpose of the
> "sort on column" feature.

> ######################### Here's the code I have so far:
> [joe@localhost work]$ cat test18.pl
> #!/usr/bin/perl

> my $infile = "test18.in";
> open(F,"$infile") || print "cannot open $infile $!";
> my @array = <F>;
> close(F);

> numericSortCol(1,\@array);

> sub numericSortCol {
> my $col = shift;
> my $aref = shift;
> foreach my $item (sort bynum($col,@{ $aref })){ # <<<< problem
> line??

Look again at the documentation for the sort function. It takes
either a block or a function (that itself expects two arguments)
and, as the second argument, a list to be sorted. Your use of
sort doesn't seem to fit that very well and I guess if you had
used 'use warnings;' you would have been told so...

> print "item: $item\n";
> }
> }

> sub bynum {
> my $col = shift;
> @a = split(/\s+/,$a);
> @b = split(/\s+/,$b);
> $a[$col] <=> $b[$col];
> }

The simplest solution is probaly not to use a function name when
calling sort but instead a block like this:

#!/usr/bin/perl

use strict;
use warnings;

my @array = <DATA>;
numericSortCol( 1, \@array );

sub numericSortCol {
my ( $col, $aref ) = @_;

print "item: $_"
for sort { ( split /\s+/, $a )[ $col ] <=>
( split /\s+/, $b )[ $col ] } @$aref;
}

__DATA__
300 400 500 600 700
33 2337483 482 78374 4567
10 20 30 40 50
1000 1001 1002 1003 1004
9 8 7 6 5
Regards, Jens
--
\ Jens Thoms Toerring ___ jt@toerring.de
\__________________________ http://toerring.de

On Thu, 17 Jul 2008 14:29:52 -0700 (PDT) joemacbusiness@yahoo.com wrote:

j> I want a subroutine that will sort a file on any given column.
j> So I have a file like this:

j> 300 400 500 600 700
j> 33 2337483 482 78374 4567
j> 10 20 30 40 50
j> 1000 1001 1002 1003 1004
j> 9 8 7 6 5

j> And I run my numericSortCol() routine on it sorting on
j> column 1 and should get this:

j> 9 8 7 6 5
j> 10 20 30 40 50
j> 300 400 500 600 700
j> 1000 1001 1002 1003 1004
j> 33 2337483 482 78374 4567

Use the standard Unix utility `sort':

sort -n +1 FILE
(use "-k 1" instead of +1 if GNU sort is installed)

e.g. sort /etc/passwd by user ID:

sort -t: -n +2 /etc/passwd

Read the docs for `sort' for further information.

You can use Perl for this (as others have shown), but if all you want is
to sort a file by a simple key, `sort' will do just fine. It's also
likely to be much faster in most circumstances.

Ted