Problem with global text search using Regular Expressions

[Jayal]

Hi
I am trying to perform a global search in a string using regular
expressions.

consider the string

"attgccgccgccatt"
If I wish to search for the string "ccgcc" within it, if I use the
operation given below, I would get only one find

$a =~ /ccgcc/g;
$location = pos......

which would be the find match.
What do I do if I want to get both occurances?

I also tried the index() function but that does not allow use of
regular expressions. Is there any function available for this?

Thank you in advance

Jayal

 

[A. Sinan Unur]

(e-mail address removed) (Jayal) wrote in @posting.google.com:

Hi
I am trying to perform a global search in a string using regular
expressions.

consider the string

"attgccgccgccatt"
If I wish to search for the string "ccgcc" within it, if I use the
operation given below, I would get only one find

$a =~ /ccgcc/g;
$location = pos......

which would be the find match.
What do I do if I want to get both occurances?

You'll need to read up on your Perl regexes. There 'ccgcc' occurs only
once in the string above. Once a match is made, the search for a next one
does starts at the third 'g' in the source string.

That is,

my $s = 'attgccgccgccatt';

while( $s =~ /ccgcc/g ) {
print 'Position: ', pos $s, "\n";
}

will only print 9.

I remember seeing something to deal with this kind of a situation in the
Cookbook, but I cannot remember how it is done.

 

[Matt Garrish]

(e-mail address removed) (Jayal) wrote in @posting.google.com:


You'll need to read up on your Perl regexes. There 'ccgcc' occurs only
once in the string above. Once a match is made, the search for a next one
does starts at the third 'g' in the source string.

That is,

my $s = 'attgccgccgccatt';

while( $s =~ /ccgcc/g ) {
print 'Position: ', pos $s, "\n";
}

will only print 9.

I remember seeing something to deal with this kind of a situation in the
Cookbook, but I cannot remember how it is done.

You can reset pos after the match. In this case, you'd set it back two
characters to match the next ccgcc:

my $s = 'attgccgccgccatt';

while( $s =~ /ccgcc/g ) {
print 'Position: ', pos $s, "\n";
pos($s) -= 2;
}

Matt

 

[Josef Moellers]

You can reset pos after the match. In this case, you'd set it back two
characters

I'd suggest length('ccgcc') - 1
then e.g. a search for 'cccc' in 'cccccc' would produce the "expected" 3
matches.
Agreed, it wouldn't produce any more matches in the original case and
would take slightly longer, but for the sake of genericity ...

 

[Anno Siegel]

(e-mail address removed) (Jayal) wrote in @posting.google.com:


You'll need to read up on your Perl regexes. There 'ccgcc' occurs only
once in the string above. Once a match is made, the search for a next one
does starts at the third 'g' in the source string.

That is,

my $s = 'attgccgccgccatt';

while( $s =~ /ccgcc/g ) {
print 'Position: ', pos $s, "\n";
}

will only print 9.

I remember seeing something to deal with this kind of a situation in the
Cookbook, but I cannot remember how it is done.

It probably uses lookahead in place of a plain match. Lookahead
doesn't consume the characters it looks at, so it can detect overlapping
matches. It is still possible to capture the match "inside" the
lookahead. So

print "$\n" for $s =~ /(?=(ccgcc))/g;

prints "ccgcc" twice.

Anno

 

[Anno Siegel]

[...]

print "$\n" for $s =~ /(?=(ccgcc))/g;

prints "ccgcc" twice.

Anno

I presume you mean:

print "$_\n" for $s =~ /(?=(ccgcc))/g;

Yes, that's what I meant.

Anno

 

[A. Sinan Unur]

(e-mail address removed)-berlin.de (Anno Siegel) wrote in

It probably uses lookahead in place of a plain match.

Thanks Anno. Should've thought of that but I rarely use anything other than
the most basic regexes. Thanks also to Matt for reminding me that pos
returns an lvalue.

Sinan.

 

[Anno Siegel]

(e-mail address removed)-berlin.de (Anno Siegel) wrote in



Thanks Anno. Should've thought of that but I rarely use anything other than
the most basic regexes.

....which is as it should be. Only a small regex is a good regex

Anno

 

[Charles DeRykus]

Hi
I am trying to perform a global search in a string using regular
expressions.

consider the string

"attgccgccgccatt"
If I wish to search for the string "ccgcc" within it, if I use the
operation given below, I would get only one find

$a =~ /ccgcc/g;
$location = pos......

which would be the find match.
What do I do if I want to get both occurances?

I also tried the index() function but that does not allow use of
regular expressions. Is there any function available for this?

With a slight mod to Anno's solution, you could even get
match pos's:

my $pos = 0;
while ( $s =~ /(.+?)(?=(ccgcc))/g ) {
print "pos=", ($pos += length $1), " match=$2\n";

}

pos=4 match=ccgcc
pos=7 match=ccgcc


hth,

 

[Charles DeRykus]

With a slight mod to Anno's solution, you could even get
match pos's:

my $pos = 0;
while ( $s =~ /(.+?)(?=(ccgcc))/g ) {
print "pos=", ($pos += length $1), " match=$2\n";

}

pos=4 match=ccgcc
pos=7 match=ccgcc

Needs a mod to the mod:

# mod to handle match string at pos 0
my $pos = 0;
while ( $s =~ /(^|.+?)(?=(ccgcc))/g ) {
print "pos=", ($pos += length $1), " match=$2\n";
}