Reading into a hash from regex

songahji has asked for the wisdom of the Perl Monks concerning the following question:

Lets say I have a file containing

--------------------
ASCII=i
Character=i
Decimal=970
Hexidecimal=3CA
Entity name=
Decription=i with an umlaut
--------------------
ASCII='
Character='
Decimal=39
Hexidecimal=27
Entity name=
Decription=quote
---------------------
ASCII=<
Character=<
Decimal=60
Hexidecimal=3C
Entity name=lt
Decription=less than
--------------------
.... truncated
[download]

Next, a hash need to be populated with the value of 'ASCII' and value of 'Character'. Since the value of 'Character' are unique, it will be the hash key. So far this is what I came up

open IN, "$myfile";
{
    local $/;
    $str = <IN>;
}
close IN;
%hash = reverse ($str =~ m/ASCII\=(.*?)\s+Character\=(.*?)\s+/g);
[download]

Are there other ways to do this better?

Greets,
Hanny J

Comment on Reading into a hash from regex Select or Download Code

Replies are listed 'Best First'.
Re: Reading into a hash from regex by holli (Abbot) on May 16, 2005 at 20:38 UTC
maybe, `use strict; use warnings; $/="--------------------"; my %h = (); open IN, "<", "c:/pm217.txt" or die $!; while (<IN>) { if ( /ASCII=(.+?)\nCharacter=(.+?)\n/ms ) { $h{$2} = $1; } } close IN;` [download] holli, /regexed monk/	[reply] [d/l]
Re: Reading into a hash from regex by jdporter (Paladin) on May 16, 2005 at 20:42 UTC
Or: `my %h = do { local $/ = "--------------------"; local @ARGV = "c:/pm217.txt"; map { ( /Character=(.)/, /ASCII=(.)/ ) } <> };` [download] Or: `my %h; { local $/; # sluurp local @ARGV = "c:/pm217.txt"; @h{ /Character=(.)/g } = /ASCII=(.)/g for <>; }` [download]	[reply] [d/l] [select]
Re: Reading into a hash from regex by TedPride (Priest) on May 16, 2005 at 21:05 UTC
`use strict; use warnings; my ($handle, %hash, $a, $c); my $myfile = 'whatever.dat'; open $handle, $myfile; <$handle>; while ($_ = <$handle>) { $a = substr($_, 6, 1); $c = substr(<$handle>, 10, 1); <$handle> for (0..4); $hash{$c} = $a; }` [download]	[reply] [d/l]
Re: Reading into a hash from regex by tlm (Prior) on May 17, 2005 at 00:04 UTC
You already got some good alternative snippets that fix these errors, but I thought it would be helpful to point them out explicitly. Namely, your regexp will give the wrong results when what immediately follows the `=` matches `\s`. Therefore you should replace the `.*?` in the captures with `.+?`. Also, you should add the `/s` modifier to the regexp if you want the `.` in the captures to match `\n`. the lowliest monk	[reply]
Re^2: Reading into a hash from regex by songahji (Friar) on May 17, 2005 at 13:42 UTC
Well spotted. Thx, Hanny J	[reply]
Re: Reading into a hash from regex by strictvars (Sexton) on May 17, 2005 at 19:48 UTC
Maybe with a LISP style `BEGIN{ open IN, $myfile or die "can't open $!"; local $/; $char{ ( /Character=(.)/ )[0] } = $hex{ ( /Hexidecimal=(.)/ )[0] + } = { map{ /(.?)=(.)/ } split /\n/ } for split /\n-+\n/, <IN>; }` [download]	[reply] [d/l]
Re: Reading into a hash from regex by Roger (Parson) on May 17, 2005 at 13:23 UTC
Here's my try... I make an assumption that either ASCII or Character could be the first line of the section, therefore some re-arrangement is needed: use Data::Dumper; my %hash; my @n = map { /(Character\|ASCII)\s?=\s?(\S+)/; $1 ? [$1,$2] : () } < +DATA>; while (my ($a, $b) = splice @n, 0, 2) { if ($a->[0] eq 'ASCII') { $hash{$b->[1]} = $a->[1]; } else { $hash{$a->[1]} = $b->[1]; } } print Dumper(\%hash); __DATA__ -------------------- ASCII=1 Character=2 Decimal=970 Hexidecimal=3CA Entity name= Decription=i with an umlaut -------------------- Character=4 ASCII=3 Decimal=39 Hexidecimal=27 Entity name= Decription=quote --------------------- ASCII=< Character=< Decimal=60 Hexidecimal=3C Entity name=lt Decription=less than [download]	[reply] [d/l]