The input file looks like this:

examples/rows_of_file.txt

A 1
B 2
A 2
C 3
B 2
D 4
C 1 

The expected output looks like this:

A 1
B 2
C 3
D 4

What we need to notice is that only the first characters count. The first appearance of each character as the head of each line will determine which line is included in the final result.

We need to find a way to remember which chracters have been already seen as first-characters of a line. If this is the first appearance of a given character, inlcude the line in the output file. If the character was already seen, we skip this line.

It would be easier to do this using a hash, but at this point in the course we have not learned hashes yet. So we need to settle using an array.

For that we need a mapping from each character that can be the first character of a line to a unique location in an array. Luckily the ord function returns the ASCII code of each ASCII character. We can use that for the mapping.

The first 32 characterts of the ASCII table are not printable characters, and if we assume we only have letters as the fist character then many of the numbers between 0 and 127 won't be used. This means if we use the ASCII code as index in the array, we'll have a sparse array with only some of the fields having a value.

As the total number of potential values is only 128, in most situation we can get away with this waste.

examples/unique_rows_with_array.pl

use strict;
use warnings;

my ($infile, $outfile) = @ARGV;
die "USAGE: $0 INFILE OUTFILE\n" if not $outfile;

open my $in,  '<', $infile  or die "Could not open '$infile': $!";
open my $out, '>', $outfile or die "Could not open '$outfile': $!";

my @seen;
while (my $line = <$in>) {
    my $chr = substr $line, 0, 1;
    my $ascii = ord $chr;
    if ($seen[$ascii]) {
        next;
    }
    $seen[$ascii] = 1;
    print $out $line;
}

First we expect the name of the input file and the name of the output file on the command line.

Then we open them for reading and writing respectively.

Then we declare an array, cleverly named @seen to indicate if a given character has been already seen as the leading character of a line. It starts out empty as we have not seen any charcter yet.

As we go over the lines of the input file using a while loop we extract the first character using substr.

Then using the ord function we get the ASCII code of that character.

If the corresponding value in the @seen array is already True we skip the rest of the block by calling next.

Otherwise we set that field to 1. It can be any arbitrary value, we are only interested in True-ness of that value.

Finally we print the current line.

Comments

Interesting solution, but what is the point of the "ord" function? Can you not compare the value of "$chr"?

Can you explain why just the first character as the title says "display unique rows of a file"? I would think the entire row would be considered in that case. I do realize there are cases where just the first character or "word" is important like a config file where you can override settings.

I'd like to see a file check added to this line as follows:

die "USAGE: $0 INFILE OUTFILE\n" if not $outfile;

die "USAGE: $0 INFILE OUTFILE\n" if not (-s $infile && $outfile);