How to use local * on reference arguments? - perl

This question is related to the last point of Item 46 in Effective Perl Programming.
I tested out this function, which allows you to pass array references, but access them as local arrays:
use strict;
sub max_v_local {
local ( *a, *b ) = #_;
my $n = #a > #b ? #a : #b;
my #result;
for ( my $i = 0 ; $i < $n ; $i++ ) {
push #result, $a[$i] > $b[$i] ? $a[$i] : $b[$i];
}
#result;
}
But I got the following errors unless I don't use strict:
Variable "#a" is not imported
Variable "#b" is not imported
Global symbol "#a" requires explicit package name
Global symbol "#b" requires explicit package name
Is there a way to do this with strict?
Update
Some further background. The above subroutine was a refinement of what follows. The subroutine takes 2 arrayrefs, but using the arrayrefs in the subroutine can get messy. The above code will probably be faster and is more readable since it enables you to access the arrayrefs as local arrays.
sub max_v {
my ( $a, $b ) = #_;
my $n = #$a > #$b ? #$a : #$b; # no. of items
my #result;
for ( my $i = 0 ; $i < $n ; $i++ ) {
push #result, $$a[$i] > $$b[$i] ? $$a[$i] : $$b[$i];
}
#result;
}
I haven't paid much attention to globs prior to this, so I'm having a look at them now. Turns out that they're not as complicated as I thought.

Yes. Declare them with our:
use strict;
sub max_v_local {
local ( *x, *y ) = #_;
our (#x, #y);
my $n = #x > #y ? #x : #y;
my #result;
for ( my $i = 0 ; $i < $n ; $i++ ) {
push #result, $x[$i] > $y[$i] ? $x[$i] : $y[$i];
}
#result;
}
(It's generally not a good idea to use variables named a or b for anything except sort.)

As cjm mentions, you need to declare the variables with our (or use fully qualified names).
Now for a few tips. First off, assigning from #_ into a glob unchecked is a bit of a gamble. I would write the line like this:
our (#a, #b);
local (*a, *b) = map \#$_ => #_;
That way, you have ensured that the only things passed to your subroutine are actually array references. Perl will throw an error if the value is not an array reference. You can of course write a verbose check if you want a more detailed message:
ref eq 'ARRAY' or die "..." for #_;
our (#a, #b);
local (*a, *b) = #_;

You want to use the package variables #a and #b, so you want to use our. It's almost a per-variable no strict "vars";, and it's lexically scoped.
sub max_v_local {
local ( *a, *b ) = #_;
our ( #a, #b );
...
}
But there is a problem in your algorithm. You're comparing elements of the longer array to elements that don't exist. That will give warnings and give the wrong result for negative values. Fix:
sub max_v_local {
local ( *a, *b ) = #_;
our ( #a, #b );
my $n = #a < #b ? #a : #b;
return
( map { $a[$_] > $b[$_] ? $a[$_] : $b[$_] } 0..$n-1 ),
#a[ #b .. $#a ],
#b[ #a .. $#b ];
}
Also, it's odd to take references and return a list, though. You might want to return an array reference. (return [ ... ];)

Think it might be because you haven't explicitly declared your variable #a and #b

Related

How to overload operator in non-class package?

In my situation I don't need warnings Use of uninitialized value in string while comparing string equality. So I tought that instead silencing all such warnings in the scope with no warnings 'uninitialized' would be better to overload eq-operator with my own subroutine, like:
use overload 'eq' => \&undefined_equal;
sub undefined_equal {
my ( $left, $right ) = #_;
no warnings 'uninitialized';
if ( $left eq $right ) {
return 1;
}
return 0;
}
Of course, overloading does not work, because according to the docs, overload is meant to use with classes, but I have plain procedural packages.
So I did try with overloading built-in functions, like:
package REeq;
use strict; use warnings; use 5.014;
BEGIN {
use Exporter ();
#REeq::ISA = qw( Exporter );
#REeq::EXPORT = qw( eq );
}
sub eq {
my ( $left, $right ) = #_;
no warnings 'uninitialized';
if ( $left CORE::eq $right ) {
return 1;
}
return 0;
}
1;
I can call my eq but can't use it as operator.
I need it because I want instead
if ( defined $some_scalar && $some_scalar eq 'literal string' ){
....
}
to use just
if ( $some_scalar eq 'literal string' ){
....
}
How could I achieve my goal?
Changing the behaviour of eq is possible, but it requires writing an XS modules that creates an op checker that replaces the code perl executes for the eq ops in scope. This is the approach used by no autovivification;, for example.
Seems you can - I haven't tried this but the perl monks have
sure, if you see it that way... you just have to bless your variable,
just like you did with your Number-package.
use overload ...;
my $n = bless {number => 23}, "main";
print $n >> 2;
i think that's not what you want, just wanted to make clear that it's
not a problem of the package name but that you must have a blessed
object.
Edit: taking zdim's onboard...
use strict;
use warnings;
use overload 'eq' => \&undefined_equal;
sub undefined_equal {
my ( $left, $right ) = #_;
no warnings 'uninitialized';
if ( ${$left} eq $right ) {
return 1;
}
return 0;
}
my $a = "abcde";
my $n = bless \$a, "main";
print "a eq undef -->";
print $a eq undef;
print "<--\nn eq undef -->";
print $n eq undef;
print "<--\n";
which gives
$ perl overload.pl
Use of uninitialized value in string eq at overload.pl line 20.
a eq undef --><--
n eq undef -->0<--
Don't forget the double $$ in the sub or you disappear into recursion. And the scalar reference for bless as you can only bless references, it seems
It still has a bless but hey

Unexpected results for high order function

I have a higher order function that maps even position values in an array:
sub map_even(&#) {
my $block = shift;
my #res;
for $i (0..$#_) {
push #res, $i%2 ? $_[$i] : &$block($_[$i]);
}
#res;
}
print map_even {$_*$_} 1,2,3,4;
I am expecting the output to be 14316, but the actual output is
0204
Why does this happen and how can I fix this? And is there any improvement can be done to the code?
In your anonymous function you have to access first input argument via $_[0] (hint: #_ array).
use strict;
use warnings;
sub map_even(&#) {
my $block = shift;
my #res;
for my $i (0..$#_) {
push #res, $i%2 ? $block->($_[$i]) : $_[$i];
}
#res;
}
print join ",", map_even {$_[0]*$_[0]} 1,2,3,4;
output
1,4,3,16
Using $_,
sub map_even(&#) {
my $block = shift;
my #res;
for my $i (0..$#_) {
push #res, $i%2 ? $block->() : $_ for $_[$i];
# or
# local $_ = $_[$i];
# push #res, $i%2 ? $block->() : $_;
}
#res;
}
print join ",", map_even {$_*$_} 1,2,3,4;
In your map_even block, you use the special $_ variable. However, you have to set it inside your loop:
local $_ = $_[$i];
... $block->();
The $_ is a global variable and can be temporarily overridden with the local operator. The $_ has nothing to do with subroutine arguments.
About aliasing: Perls for, map and grep mostly alias $_ to the current element as a performance hack, not because this behavior would be particularly desirable. In order to perform an alias, you should localize the whole *_ typeglob which contains the $_ variable and then assign a scalar reference of the alias target to the glob:
local *_ = \$_[$i];
I would solve this one of two ways.
First, by using List::Utils's pairmap:
use strict;
use warnings;
use List::Util qw(pairmap);
my #x = (1 .. 4);
my #result = pairmap {$a, $b**2} #x;
print "#result\n";
Or more simply, by just using the indexes:
use strict;
use warnings;
my #x = (1 .. 4);
my #result = map {$_ % 2 ? $x[$_] ** 2 : $x[$_]} (0..$#x);
print "#result\n";
However, if you really wanted a new sub, I'd just setup a flip-flop:
use strict;
use warnings;
sub map_even(&#) {
my $block = shift;
my $even = 1;
map {($even ^= 1) ? $block->() : $_} #_;
}
print join " ", map_even {$_*$_} 1,2,3,4;
All output:
1 4 3 16

How can I get this basic Perl sub program that sorts to work properly?

I am brand new to Perl. Can someone help me out and give me a tip or a solution on how to get this sorting sub program to work. I know it has something to do with how arrays are passed to sub programs. I searched online and did not find an answer that I was satisfied with... I also like the suggestions the helpful S.O. users give me too. I would like to have the program print the sorted array in the main sub program. Currently, it is printing the elements of the array #a in original order. I want the sub program to modify the array so when I print the array it is in sorted order. Any suggestions are appreciated. Of course, I want to see the simplest way to fix this.
sub sort {
my #array = #_;
my $i;
my $j;
my $iMin;
for ( $i = 0; $i < #_ - 1; $i++ ) {
$iMin = $i;
for ( $j = $i + 1; $j < #_; $j++ ) {
if ( $array[$j] < $array[$iMin] ) {
$iMin = $j;
}
}
if ( $iMin != $i ) {
my $temp = $array[$i];
$array[$i] = $array[$iMin];
$array[$iMin] = $temp;
}
}
}
Then call from a main sub program:
sub main {
my #a = (-23,3,234,-45,0,32,12,54,-10000,1);
&sort(#a);
my $i;
for ( $i = 0; $i < #a; $i++ ) {
print "$a[$i]\n";
}
}
main;
When your sub does the following assignment my #array = #_, it is creating a copy of the passed contents. Therefore any modifications to the values of #array will not effect #a outside your subroutine.
Following the clarification that this is just a personal learning exercise, there are two solutions.
1) You can return the sorted array and assign it to your original variable
sub mysort {
my #array = #_;
...
return #array;
}
#a = mysort(#a)
2) Or you can pass a reference to the array, and work on the reference:
sub mysort {
my $arrayref = shift;
...
}
mysort(\#a)
Also, it's probably a good idea to not use a sub named sort since that's that's a builtin function. Duplicating your code using perl's sort:
#a = sort {$a <=> $b} #a;
Also, the for loops inside your sub should be rewritten to utilize the last index of an #array, which is written as $#array, and the range operator .. which is useful for incrementors :
for ( my $j = $i + 1; $j <= $#array; $j++ ) {
# Or simpler:
for my $j ($i+1 .. $#array) {
And finally, because you're new, I should pass on that all your scripts should start with use strict; and use warnings;. For reasons why: Why use strict and warnings?
With very few, rare exceptions the simplest (and easiest) way to sort stuff in perl is simply to use the sort builtin.
sort takes an optional argument, either a block or a subname, which can be used to control how sort evaluates which of the two elements it is comparing at any given moment is greater.
See sort on perldoc for further information.
If you require a "natural" sort function, where you get the sequence 0, 1, 2, 3, ... instead of 0, 1, 10, 11, 12, 2, 21, 22, 3, ..., then use the perl module Sort::Naturally which is available on CPAN (and commonly available as a package on most distros).
In your case, if you need a pure numeric sort, the following will be quite sufficient:
use Sort::Naturally; #Assuming Sort::Naturally is installed
sub main {
my #a = (-23,3,234,-45,0,32,12,54,-10000,1);
#Choose one of the following
#a = sort #a; #Sort in "ASCII" ascending order
#a = sort { $b cmp $a } #a; #Sort in reverse of the above
#a = nsort #a; #Sort in "natural" order
#a = sort { ncmp($b, $a) } #a; #Reverse of the above
print "$_\n" foreach #a; #To see what you actually got
}
It is also worth mentioning the use sort 'stable'; pragma which can be used to ensure that sorting occurs using a stable algorithm, meaning that elements which are equal will not be rearranged relative to one another.
As a bonus, you should be aware that sort can be used to sort data structures as well as simple scalars:
#Assume #a is an array of hashes
#a = sort { $a->{name} cmp $b->{name} } #; #Sort #a by name key
#Sort #a by name in ascending order and date in descending order
#a = sort { $a->{name} cmp $b->{name} || $b->{date} cmp $a->{date} } #a;
#Assume #a is an array of arrays
#Sort #a by the 2nd element of the arrays it contains
#a = sort { $a->[1] cmp $b->[1] } #a;
#Assume #a is an array of VERY LONG strings
#Sort #a alphanumerically, but only care about
#the first 1,000 characters of each string
#a = sort { substr($a, 0, 1000) cmp substr($b, 0, 1000) } #a;
#Assume we want to "sort" an array without modifying it:
#Yes, the names here are confusing. See below.
my #idxs = sort { $a[$a] cmp $a[$b] } (0..$#a);
print "$a[$_]\n" foreach #idxs;
##idxs contains the indexes to #a, in the order they would have
#to be read from #a in order to get a sorted version of #a
As a final note, please remember that $a and $b are special variables in perl, which are pre-populated in the context of a sorting sub or sort block; the upshot is that if you're working with sort you can always expect $a and $b to contain the next two elements being compared, and should use them accordingly, but do NOT do my $a;, e.g., or use variables with either name in non-sort-related stuff. This also means that naming things %a or #a, or %b or #b, can be confusing -- see the final section of my example above.

perl: iterate over a typeglob

Given a typeglob, how can I find which types are actually defined?
In my application, we user PERL as a simple configuration format.
I'd like to require() the user config file, then be able to see which variables are defined, as well as what types they are.
Code: (questionable quality advisory)
#!/usr/bin/env perl
use strict;
use warnings;
my %before = %main::;
require "/path/to/my.config";
my %after = %main::;
foreach my $key (sort keys %after) {
next if exists $before{$symbol};
local *myglob = $after{$symbol};
#the SCALAR glob is always defined, so we check the value instead
if ( defined ${ *myglob{SCALAR} } ) {
my $val = ${ *myglob{SCALAR} };
print "\$$symbol = '".$val."'\n" ;
}
if ( defined *myglob{ARRAY} ) {
my #val = #{ *myglob{ARRAY} };
print "\#$symbol = ( '". join("', '", #val) . "' )\n" ;
}
if ( defined *myglob{HASH} ) {
my %val = %{ *myglob{HASH} };
print "\%$symbol = ( ";
while( my ($key, $val) = each %val ) {
print "$key=>'$val', ";
}
print ")\n" ;
}
}
my.config:
#A = ( a, b, c );
%B = ( b=>'bee' );
$C = 'see';
output:
#A = ( 'a', 'b', 'c' )
%B = ( b=>'bee', )
$C = 'see'
$_<my.config = 'my.config'
In the fully general case, you can't do what you want thanks to the following excerpt from perlref:
*foo{THING} returns undef if that particular THING hasn't been used yet, except in the case of scalars. *foo{SCALAR} returns a reference to an anonymous scalar if $foo hasn't been used yet. This might change in a future release.
But if you're willing to accept the restriction that any scalar must have a defined value to be detected, then you might use code such as
#! /usr/bin/perl
use strict;
use warnings;
open my $fh, "<", \$_; # get DynaLoader out of the way
my %before = %main::;
require "my.config";
my %after = %main::;
foreach my $name (sort keys %after) {
unless (exists $before{$name}) {
no strict 'refs';
my $glob = $after{$name};
print "\$$name\n" if defined ${ *{$glob}{SCALAR} };
print "\#$name\n" if defined *{$glob}{ARRAY};
print "%$name\n" if defined *{$glob}{HASH};
print "&$name\n" if defined *{$glob}{CODE};
print "$name (format)\n" if defined *{$glob}{FORMAT};
print "$name (filehandle)\n" if defined *{$glob}{IO};
}
}
will get you there.
With my.config of
$JACKPOT = 3_756_788;
$YOU_CANT_SEE_ME = undef;
#OPTIONS = qw/ apple cherries bar orange lemon /;
%CREDITS = (1 => 1, 5 => 6, 10 => 15);
sub is_jackpot {
local $" = ""; # " fix Stack Overflow highlighting
"#_[0,1,2]" eq "barbarbar";
}
open FH, "<", \$JACKPOT;
format WinMessage =
You win!
.
the output is
%CREDITS
FH (filehandle)
$JACKPOT
#OPTIONS
WinMessage (format)
&is_jackpot
Printing the names takes a little work, but we can use the Data::Dumper module to take part of the burden. The front matter is similar:
#! /usr/bin/perl
use warnings;
use strict;
use Data::Dumper;
sub _dump {
my($ref) = #_;
local $Data::Dumper::Indent = 0;
local $Data::Dumper::Terse = 1;
scalar Dumper $ref;
}
open my $fh, "<", \$_; # get DynaLoader out of the way
my %before = %main::;
require "my.config";
my %after = %main::;
We need to dump the various slots slightly differently and in each case remove the trappings of references:
my %dump = (
SCALAR => sub {
my($ref,$name) = #_;
return unless defined $$ref;
"\$$name = " . substr _dump($ref), 1;
},
ARRAY => sub {
my($ref,$name) = #_;
return unless defined $ref;
for ("\#$name = " . _dump $ref) {
s/= \[/= (/;
s/\]$/)/;
return $_;
}
},
HASH => sub {
my($ref,$name) = #_;
return unless defined $ref;
for ("%$name = " . _dump $ref) {
s/= \{/= (/;
s/\}$/)/;
return $_;
}
},
);
Finally, we loop over the set-difference between %before and %after:
foreach my $name (sort keys %after) {
unless (exists $before{$name}) {
no strict 'refs';
my $glob = $after{$name};
foreach my $slot (keys %dump) {
my $var = $dump{$slot}(*{$glob}{$slot},$name);
print $var, "\n" if defined $var;
}
}
}
Using the my.config from your question, the output is
$ ./prog.pl
#A = ('a','b','c')
%B = ('b' => 'bee')
$C = 'see'
Working code using a CPAN module that gets some of the hair out of the way, Package::Stash. As noted in my comment to gbacon's answer, this is blind to the config file doing $someval = undef but that seems to be unavoidable, and at least the other cases are caught. It also limits itself to the SCALAR, ARRAY, HASH, CODE, and IO types -- getting GLOB and FORMAT is possible but it makes the code less pretty and also creates noise in the output :)
#!perl
use strict;
use warnings;
use Package::Stash;
sub all_vars_in {
my ($package) = #_;
my #ret;
my $stash = Package::Stash->new($package);
for my $sym ($stash->list_all_package_symbols) {
for my $sigil (qw($ # % &), '') {
my $fullsym = "$sigil$sym";
push #ret, $fullsym if $stash->has_package_symbol($fullsym);
}
}
#ret;
}
my %before;
$before{$_} ++ for all_vars_in('main');
require "my.config";
for my $var (all_vars_in('main')) {
print "$var\n" unless exists $before{$var};
}
Beginning in 5.010, you can distinguish whether a SCALAR exists using the B introspection module; see Detecting declared package variables in perl
Update: example copied from that answer:
# package main;
our $f;
sub f {}
sub g {}
use B;
use 5.010;
if ( ${ B::svref_2object(\*f)->SV } ) {
say "f: Thar be a scalar tharrr!";
}
if ( ${ B::svref_2object(\*g)->SV } ) {
say "g: Thar be a scalar tharrr!";
}
1;
UPDATE:
gbacon is right. *glob{SCALAR} is defined.
Here is the output I get using your code:
Name "main::glob" used only once:
possible typo at
test_glob_foo_thing.pl line 13.
'FOO1' (SCALAR)
'FOO1' (GLOB)
'FOO2' (SCALAR)
'FOO2' (GLOB)
'_<my.config' (SCALAR)
'_<my.config' (GLOB)
This is despite FOO2 being defined as a hash, but not as a scalar.
ORIGINAL ANSWER:
If I understand you correctly, you simply need to use the defined built-in.
#!/usr/bin/env perl
use strict;
use warnings;
my %before = %main::;
require "/path/to/my.config";
my %after = %main::;
foreach my $key (sort keys %after) {
if (not exists $before{$key}) {
if(defined($after{$key}){
my $val = $after{$key};
my $what = ref($val);
print "'$key' ($what)\n";
}
}
}
I hate to ask, but instead of messing around with typeglobs, why not switch to a real configuration format? e.g. check out Config::Simple and YAML.
I wouldn't recommend messing around with typeglobs and symbol tables in normal cases (some CPAN modules do that, but only at the bottom levels of large systems - e.g. Moose in the lowest levels of Class::MOP). Perl gives you a lot of rope to work with, but that rope is also quite happy to self-noosify and self-tie-around-your-neck if you're not careful :)
See also: How do you manage configuration files in Perl?
no strict 'refs';
my $func_name = 'myfunc';
*{$func_name}{CODE}()
use strict 'refs';
If you don't mind parsing Data::Dump output, you could use it to tease out the differences.
use strict;
use warnings;
use Data::Dump qw{ dump };
my %before = %main::;
require "my.config";
my %after = %main::;
foreach my $key ( sort keys %after ) {
if ( not exists $before{$key} ) {
my $glob = $after{$key};
print "'$key' " . dump( $glob) . "\n";
}
}
Using this code with the following config file:
$FOO1 = 3;
$FOO2 = 'my_scalar';
%FOO2 = ( a=>'b', c=>'d' );
#FOO3 = ( 1 .. 5);
$FOO4 = [ 1 .. 5 ];
I believe that this output provides enough information to be able to figure out which parts of each type glob are defined:
'FOO1' do {
my $a = *main::FOO1;
$a = \3;
$a;
}
'FOO2' do {
my $a = *main::FOO2;
$a = \"my_scalar";
$a = { a => "b", c => "d" };
$a;
}
'FOO3' do {
my $a = *main::FOO3;
$a = [1 .. 5];
$a;
}
'FOO4' do {
my $a = *main::FOO4;
$a = \[1 .. 5];
$a;
}
'_<my.config' do {
my $a = *main::_<my.config;
$a = \"my.config";
$a;
}

How do I determine the longest similar portion of several strings?

As per the title, I'm trying to find a way to programmatically determine the longest portion of similarity between several strings.
Example:
file:///home/gms8994/Music/t.A.T.u./
file:///home/gms8994/Music/nina%20sky/
file:///home/gms8994/Music/A%20Perfect%20Circle/
Ideally, I'd get back file:///home/gms8994/Music/, because that's the longest portion that's common for all 3 strings.
Specifically, I'm looking for a Perl solution, but a solution in any language (or even pseudo-language) would suffice.
From the comments: yes, only at the beginning; but there is the possibility of having some other entry in the list, which would be ignored for this question.
Edit: I'm sorry for mistake. My pity that I overseen that using my variable inside countit(x, q{}) is big mistake. This string is evaluated inside Benchmark module and #str was empty there. This solution is not as fast as I presented. See correction below. I'm sorry again.
Perl can be fast:
use strict;
use warnings;
package LCP;
sub LCP {
return '' unless #_;
return $_[0] if #_ == 1;
my $i = 0;
my $first = shift;
my $min_length = length($first);
foreach (#_) {
$min_length = length($_) if length($_) < $min_length;
}
INDEX: foreach my $ch ( split //, $first ) {
last INDEX unless $i < $min_length;
foreach my $string (#_) {
last INDEX if substr($string, $i, 1) ne $ch;
}
}
continue { $i++ }
return substr $first, 0, $i;
}
# Roy's implementation
sub LCP2 {
return '' unless #_;
my $prefix = shift;
for (#_) {
chop $prefix while (! /^\Q$prefix\E/);
}
return $prefix;
}
1;
Test suite:
#!/usr/bin/env perl
use strict;
use warnings;
Test::LCP->runtests;
package Test::LCP;
use base 'Test::Class';
use Test::More;
use Benchmark qw(:all :hireswallclock);
sub test_use : Test(startup => 1) {
use_ok('LCP');
}
sub test_lcp : Test(6) {
is( LCP::LCP(), '', 'Without parameters' );
is( LCP::LCP('abc'), 'abc', 'One parameter' );
is( LCP::LCP( 'abc', 'xyz' ), '', 'None of common prefix' );
is( LCP::LCP( 'abcdefgh', ('abcdefgh') x 15, 'abcdxyz' ),
'abcd', 'Some common prefix' );
my #str = map { chomp; $_ } <DATA>;
is( LCP::LCP(#str),
'file:///home/gms8994/Music/', 'Test data prefix' );
is( LCP::LCP2(#str),
'file:///home/gms8994/Music/', 'Test data prefix by LCP2' );
my $t = countit( 1, sub{LCP::LCP(#str)} );
diag("LCP: ${\($t->iters)} iterations took ${\(timestr($t))}");
$t = countit( 1, sub{LCP::LCP2(#str)} );
diag("LCP2: ${\($t->iters)} iterations took ${\(timestr($t))}");
}
__DATA__
file:///home/gms8994/Music/t.A.T.u./
file:///home/gms8994/Music/nina%20sky/
file:///home/gms8994/Music/A%20Perfect%20Circle/
Test suite result:
1..7
ok 1 - use LCP;
ok 2 - Without parameters
ok 3 - One parameter
ok 4 - None of common prefix
ok 5 - Some common prefix
ok 6 - Test data prefix
ok 7 - Test data prefix by LCP2
# LCP: 22635 iterations took 1.09948 wallclock secs ( 1.09 usr + 0.00 sys = 1.09 CPU) # 20766.06/s (n=22635)
# LCP2: 17919 iterations took 1.06787 wallclock secs ( 1.07 usr + 0.00 sys = 1.07 CPU) # 16746.73/s (n=17919)
That means that pure Perl solution using substr is about 20% faster than Roy's solution at your test case and one prefix finding takes about 50us. There is not necessary using XS unless your data or performance expectations are bigger.
The reference given already by Brett Daniel for the Wikipedia entry on "Longest common substring problem" is very good general reference (with pseudocode) for your question as stated. However, the algorithm can be exponential. And it looks like you might actually want an algorithm for longest common prefix which is a much simpler algorithm.
Here's the one I use for longest common prefix (and a ref to original URL):
use strict; use warnings;
sub longest_common_prefix {
# longest_common_prefix( $|# ): returns $
# URLref: http://linux.seindal.dk/2005/09/09/longest-common-prefix-in-perl
# find longest common prefix of scalar list
my $prefix = shift;
for (#_) {
chop $prefix while (! /^\Q$prefix\E/);
}
return $prefix;
}
my #str = map {chomp; $_} <DATA>;
print longest_common_prefix(#ARGV), "\n";
__DATA__
file:///home/gms8994/Music/t.A.T.u./
file:///home/gms8994/Music/nina%20sky/
file:///home/gms8994/Music/A%20Perfect%20Circle/
If you truly want a LCSS implementation, refer to these discussions (Longest Common Substring and Longest Common Subsequence) at PerlMonks.org. Tree::Suffix would probably be the best general solution for you and implements, to my knowledge, the best algorithm. Unfortunately recent builds are broken. But, a working subroutine does exist within the discussions referenced on PerlMonks in this post by Limbic~Region (reproduced here with your data).
#URLref: http://www.perlmonks.org/?node_id=549876
#by Limbic~Region
use Algorithm::Loops 'NestedLoops';
use List::Util 'reduce';
use strict; use warnings;
sub LCS{
my #str = #_;
my #pos;
for my $i (0 .. $#str) {
my $line = $str[$i];
for (0 .. length($line) - 1) {
my $char= substr($line, $_, 1);
push #{$pos[$i]{$char}}, $_;
}
}
my $sh_str = reduce {length($a) < length($b) ? $a : $b} #str;
my %map;
CHAR:
for my $char (split //, $sh_str) {
my #loop;
for (0 .. $#pos) {
next CHAR if ! $pos[$_]{$char};
push #loop, $pos[$_]{$char};
}
my $next = NestedLoops([#loop]);
while (my #char_map = $next->()) {
my $key = join '-', #char_map;
$map{$key} = $char;
}
}
my #pile;
for my $seq (keys %map) {
push #pile, $map{$seq};
for (1 .. 2) {
my $dir = $_ % 2 ? 1 : -1;
my #offset = split /-/, $seq;
$_ += $dir for #offset;
my $next = join '-', #offset;
while (exists $map{$next}) {
$pile[-1] = $dir > 0 ?
$pile[-1] . $map{$next} : $map{$next} . $pile[-1];
$_ += $dir for #offset;
$next = join '-', #offset;
}
}
}
return reduce {length($a) > length($b) ? $a : $b} #pile;
}
my #str = map {chomp; $_} <DATA>;
print LCS(#str), "\n";
__DATA__
file:///home/gms8994/Music/t.A.T.u./
file:///home/gms8994/Music/nina%20sky/
file:///home/gms8994/Music/A%20Perfect%20Circle/
It sounds like you want the k-common substring algorithm. It is exceptionally simple to program, and a good example of dynamic programming.
My first instinct is to run a loop, taking the next character from each string, until the characters are not equal. Keep a count of what position in the string you're at and then take a substring (from any of the three strings) from 0 to the position before the characters aren't equal.
In Perl, you'll have to split up the string first into characters using something like
#array = split(//, $string);
(splitting on an empty character sets each character into its own element of the array)
Then do a loop, perhaps overall:
$n =0;
#array1 = split(//, $string1);
#array2 = split(//, $string2);
#array3 = split(//, $string3);
while($array1[$n] == $array2[$n] && $array2[$n] == $array3[$n]){
$n++;
}
$sameString = substr($string1, 0, $n); #n might have to be n-1
Or at least something along those lines. Forgive me if this doesn't work, my Perl is a little rusty.
If you google for "longest common substring" you'll get some good pointers for the general case where the sequences don't have to start at the beginning of the strings.
Eg, http://en.wikipedia.org/wiki/Longest_common_substring_problem.
Mathematica happens to have a function for this built in:
http://reference.wolfram.com/mathematica/ref/LongestCommonSubsequence.html (Note that they mean contiguous subsequence, ie, substring, which is what you want.)
If you only care about the longest common prefix then it should be much faster to just loop for i from 0 till the ith characters don't all match and return substr(s, 0, i-1).
From http://forums.macosxhints.com/showthread.php?t=33780
my #strings =
(
'file:///home/gms8994/Music/t.A.T.u./',
'file:///home/gms8994/Music/nina%20sky/',
'file:///home/gms8994/Music/A%20Perfect%20Circle/',
);
my $common_part = undef;
my $sep = chr(0); # assuming it's not used legitimately
foreach my $str ( #strings ) {
# First time through loop -- set common
# to whole
if ( !defined $common_part ) {
$common_part = $str;
next;
}
if ("$common_part$sep$str" =~ /^(.*).*$sep\1.*$/)
{
$common_part = $1;
}
}
print "Common part = $common_part\n";
Faster than above, uses perl's native binary xor function, adapted from perlmongers solution (the $+[0] didn't work for me):
sub common_suffix {
my $comm = shift #_;
while ($_ = shift #_) {
$_ = substr($_,-length($comm)) if (length($_) > length($comm));
$comm = substr($comm,-length($_)) if (length($_) < length($comm));
if (( $_ ^ $comm ) =~ /(\0*)$/) {
$comm = substr($comm, -length($1));
} else {
return undef;
}
}
return $comm;
}
sub common_prefix {
my $comm = shift #_;
while ($_ = shift #_) {
$_ = substr($_,0,length($comm)) if (length($_) > length($comm));
$comm = substr($comm,0,length($_)) if (length($_) < length($comm));
if (( $_ ^ $comm ) =~ /^(\0*)/) {
$comm = substr($comm,0,length($1));
} else {
return undef;
}
}
return $comm;
}