Re^5: Use Perl's Sort to only sort certain lines in a file?

Ok, so I took this program and and that sample input.txt, ran it, and it worked. Now I do need to levels of sort, first descending numerical and then within the equal numbers, ascending alphabetically.

I've started reading on how to use map and am a little confused, but feel like I'm close.

In this section I know that I have to declare the first character in the string such that I can then sort by it, yes? This doesn't work, but it's something close?

sub handle_section {
    my ($fh) = @_;
    my ( @entries, $line );

    while ( $line = <$fh> ) {
        last unless $line =~ m{
            ( [A-Za-z] ) # 1 first character
            ( [^(]+ )    # 2 anything except opening paren
            \s           #   space
            \(           #   opening paren
            ( \d+   )    # 3 number
        }x;

        push @entries, [ pack( 'Na*', $2, $3, $1 ), $line ];
    }

    print map  { $_->[2] }
          sort { $b->[0] cmp $a->[0] 
                    ||
                 $a->[1] cmp $b->[1]
                } @entries;

    return $line;
}
[download]

Comment on Re^5: Use Perl's Sort to only sort certain lines in a file? Select or Download Code

Replies are listed 'Best First'.
Re^6: Use Perl's Sort to only sort certain lines in a file? by Anonymous Monk on Jan 02, 2015 at 03:40 UTC
Doesn't it already sort alphabetically? :) You'll probably want to read about Schwartzian transform. Anyway, perhaps it will be clearer without `pack` `last unless $line =~ m{ ( [^(]+ ) # 1 anything except opening paren \s # space \( # opening paren ( \d+ ) # 2 number }x; push @entries, [ $2, $1, $line ]; ... print map { $_->[2] } sort { $a->[0] <=> $b->[0] # sort by $2 \|\| # or if equal $a->[1] cmp $b->[1] # sort by $1 } @entries;` [download] Our @entries is an array of arrays. Each entry (inner array) has: index 0 - stuff matched by $2 (digits) index 1 - stuff matched by $1 (beginning of $line) index 2 - the $line itself For example, one entry looks like this: `['1', 'wrestling', 'wrestling (1)']` So, we first sort by '1', then by 'wrestling'. Or we can `pack` '1' and 'wrestling' in just one string and sort by that... that's just a neat trick :) `map` simply extracts the line from entry.	[reply] [d/l] [select]
Re^7: Use Perl's Sort to only sort certain lines in a file? by grahambuck (Acolyte) on Jan 02, 2015 at 06:03 UTC
Thank you so much for taking the time to help me with this. Your code has worked perfectly. If you'll pardon my one last question (I know, I know, I feel like I can't figure anything out on my own either…), my only experience so far with writing out to a file has been in the form of this: `open(Input, '+<…') \|\| die "No such file found!"; open(Output, '>…') \|\| die "Can't find the file!"; while(<Input>) { print Output $_; }` [download] With your code, when I run it I receive a log file with all the data properly formatted. I simply cannot seem to find the right way to then print that out to my Output file. If I change `print $file;` to `return $file;` I do not receive any output so I know it's stored somewhere. There are a few more edits I need to make to the file before writing out as well. Would I write another subroutine, such as "handle_section" and place it within the primary subroutine?	[reply] [d/l] [select]
Re^8: Use Perl's Sort to only sort certain lines in a file? by Anonymous Monk on Jan 02, 2015 at 07:36 UTC
`open(Input, '+<…') \|\| die "No such file found!";` [download] Well this form of open is considered much worse then `open my $input, '+<', 'file...';` [download] First, `Input` is a global variable and those are generally to be avoided. Second, when you don't specify the mode (here `+<`) separately, Perl will treat various funny characters (whilespace, \|, etc) in filenames in a special way. Which is usually not what you want. With your code, when I run it I receive a log file with all the data properly formatted. I simply cannot seem to find the right way to then print that out to my Output file. The easiest way would be to use the shell: `perl script.pl > output.txt` [download] I'm pretty sure that works with cmd.exe too. Another easy way is to reopen STDOUT (this is where `print` prints by default - normally the terminal) at the beginning of the program: `open STDOUT, '>', 'output.txt' or die $!` [download] There are a few more edits I need to make to the file before writing out as well. Would I write another subroutine, such as "handle_section" and place it within the primary subroutine? Sure, that's one way to do it. Another one would be to write one more script and pipe the output of first script to the second (in the shell) `perl first.pl \| perl second.pl` [download] The first script prints to STDOUT and the second one reads from STDIN: `while (my $line = <STDIN>) { ... }` [download]	[reply] [d/l] [select]