Help me understand this code?

[ansabhailte]

I'm very new to Perl (coming from Bash/sed/awk) and have been going through Daily Programmer challenges to get acquainted with Perl. I saw a user solution in Perl for a date sorter (yyyy-dd-mm, mm-dd-yyyy, etc reformatted) and am having a hard time figuring out what half of this means. Can anybody walk me through understanding it?

$m{$_} = ++$i for qw[Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec];
($a, $t, $f, $b) = qw[([A-Za-z]+) \b(\d\d)\b (\d{4}) (\d{4}|\d\d)];

for (<>) {
    ($y, $m, $d) =
        /$f-$t-$t/   ? ($1, $2, $3) :
        /$t\/$t\/$t/ ? ($3, $1, $2) :
        /$t#$t#$t/   ? ($2, $1, $3) :
        /$t\*$t\*$f/ ? ($3, $2, $1) :
        /$a $t, $b/  ? ($3, $m{$1}, $2) : next;
    $y += $y < 50 ? 2000 : $y < 100 ? 1900 : 0;
    printf "%04d-%02d-%02d\n", $y, $m, $d;
}

[aaron_baugher]

I assume you're having trouble with the nested ternary operators. You could do the same thing with an if-elsif-elsif-else chain, but it wouldn't look as nice as this. This method lines up the tests and results nicely. Here's a very simple one to show what's happening:

$result =
   $a ? $b :
   $c ? $d :
   $e ? $f : undef;

So it starts left-to-right through the first ternary operator. It checks the value of $a; if that's true, it returns the value of $b, and the expression is finished. It doesn't continue on to the following ternary operators, since they're all part of the third section of the operator.

But if $a is false, then it executes that third portion, which starts with the second ternary operator, checking the value of $c. If that's true, it returns the value of $d and stops. If $c is not true, it continues on to the third ternary operator, checking the value of $e, and so on.

Adding some parens to make the precedence clearer, you'd have this:

$result =
   $a ? $b :
   ( $c ? $d :
     ( $e ? $f : undef ));

So in the case of your code, each of those ternary operator lines compares to a regex, and if it matches, it then returns the matching substrings $1, $2, and $3, as specified on the right, to $y, $m, and $d, and the statement exits, continuing with the $y+= line (which also has nested ternary operators working the same way). If none of them match, it eventually gets down to the next statement, restarting the loop because it couldn't recognize the string as having a date format.

[ansabhailte]

Ah thank you!

[ww]

Not the answer to your question (tho I will expand a bit on info sources), but the PRIMARY POINTS ILLUSTRATED BY THE CODE SHOWN ARE ... wait for them... just a little longer...

1. Use meaningful variable names.
         $a and $b violate not only that wisdom, but should also be reserved for use in comparisons
         and...
2. Don't get too lazy to type a few extra chars for clarity. Saving chars too enthusiastically will usually obfuscate your code rather than improve it.

As to info sources: perldoc func, op, module_name is an incantation that will lead you to explanatory docs, right on your own computer. For example, for the regexen, you might want to see perldoc perlretut (or one of its siblings); for printf (which is a function and thus needs a "-f" before its name, thusly) perldoc -f printf (which will give you a referal to another function instead of usefully answer the question "what does it do?" unless you are intimate with sprintf's format arguement ...but that's a whole different can of worms); and, as an exercise to the OP, we have the likes of qw (for which vist re "quote word" and friends); some lists (defined by the square brackets) and a few more.

...and welcome to Perl and to the Monastery!

[ansabhailte]

SO I've been reading perlre and perlop and I get most of it now. But I'm still confused on the $m{$1} and $m{$_} parts. What do these do?

[Anonymous Monk]

%m is a hash mapping the 3 letter month to a month number.

$m{$_} is where elements are being assigned inside a for loop.

$m{$1} is dereferencing a particular value.

-Greg

[GotToBTru]

Run it in the debugger. You can inspect the variables and see what it is doing. This loops over standard input:

for (<>) { ... }

See perlop to understand what the operators are doing.

[ansabhailte]

https://gist.github.com/coderd00d/a88d4d2da014203898af This is the input. How do I feed it into it? (I know this sounds really noobish but the debugger pops up DB1 etc and I'm not sure what it's looking for exactly).

[GotToBTru]

Perldoc is your friend. perlrun

$: perl -d prog.pl inputfile

[Marshall]

#!/usr/bin/perl;
use warnings;
use strict;

# I am not sure what data set you have?
# yyyy-mm-dd will sort in ASCII order
# with a simple sort.

# yyyy-dd-mm or mm-dd-yyyy is a bit more
# complex, eg: 2013-11-07 vs 11-07-2013.

# It is possible to detect yyyy-mm-dd
# versus yyyy-dd-mm or mm-dd-yyyy, however
# dd vs mm can be quite problematic.
# 02-03-2011 vs 03-02-2011?

# the code that you show is actually rather
# strange looking.

what are your trying to do?
show some input and expected output.

[ansabhailte]

Yeah, that confused me too.

It's a user-submitted solution to /r/dailyprogrammer.

The input is:
https://gist.github.com/coderd00d/a88d4d2da014203898af

The output would be the same list in a standard format.