next »

[Phaëton]

Okay Im trying to understand this...

 [^0-9].+?

In english,
is it:  Not zero through 9 any character except newline, zero or one times,  one or more times?

is this translation  correct?

then why is it getting two characters after the slashes.... what am i missing here...
im trying to just grab everything every http://  up until a #32

Code:

<?php

 $re = '/(http:\/\/[^0-9].+?)/';
//                     

preg_match_all( $re,
    " Vermont was the first state to pass more aggressive privacy policies
 with respect to credit reports in 1992. Maine and California are typically
 early adopters when caring for their state citizen’s privacy. 
http://www.privacyrights.org/fs/fs6a-facta.htm http://uspirg.org/uspirg.asp?id2=13649 
http://en.wikipedia.org/wiki/Fair_and_Accurate_Credit_Transactions_Act

",
    $out, PREG_PATTERN_ORDER);

var_dump($out);

?>

Code:

array
  0 =>
    array
      0 => string 'http://ww' (length=9)
      1 => string 'http://us' (length=9)
      2 => string 'http://en' (length=9)
  1 =>
    array
      0 => string 'http://ww' (length=9)
      1 => string 'http://us' (length=9)
      2 => string 'http://en' (length=9)

[Bompa]

Okay Im trying to understand this...

 [^0-9].+?

In english,
is it:  Not zero through 9 any character except newline, zero or one times,  one or more times?

is this translation  correct?

then why is it getting two characters after the slashes.... what am i missing here...
im trying to just grab everything every http://  up until a #32

Close.

$re = '/(http:\/\/[^0-9].+?)/';

That will match http:// followed by a single character that is not 0-9, followed by
by one or more of anything. The ? is either doing nothing is this case or screwing
things up. 

As quantifiers, the + usually means "one or more". The ? usually means 0 or 1 time,
but can also be used to stop the greediness of matching, like in my example below.


The block of text that you're parsing is a tough one cuz it doesn't
have clear cut endings for the match.

What I would try is to match everything from http:// to the first white space.

This perl worked in a quick test:

$string =<<here;
Vermont was the first state to pass more aggressive privacy policies
 with respect to credit reports in 1992. Maine and California are typically
 early adopters when caring for their state citizen’s privacy.
http://www.privacyrights.org/fs/fs6a-facta.htm http://uspirg.org/uspirg.asp?id2=13649
http://en.wikipedia.org/wiki/Fair_and_Accurate_Credit_Transactions_Act
here

Code:

while($string =~ /(http:\/\/.*?\s)/msg) {
  $url = $1;
  print "$url\n";
}

The * means unlimited number of times, but the ? immediately following means
"up to the first", then I have \s for any white space.

Ooops, actually, mine gets all the urls, but I didn't follow why up to a #32
cuz I did not see a #32.

If you're not confused yet, let me know, I can type more nonsense. 

Bompa

[Phaëton]

I think I get it, I just dont get preg_match_all the way

Why does it return two arrays?


print_r($out);

Code:

array
  0 =>
    array
      0 => string 'http://www.privacyrights.org/fs/fs6a-facta.htm ' (length=47)
      1 => string 'http://uspirg.org/uspirg.asp?id2=13649
' (length=39)
      2 => string 'http://en.wikipedia.org/wiki/Fair_and_Accurate_Credit_Transactions_Act
' (length=71)
  1 =>
    array
      0 => string 'http://www.privacyrights.org/fs/fs6a-facta.htm ' (length=47)
      1 => string 'http://uspirg.org/uspirg.asp?id2=13649
' (length=39)
      2 => string 'http://en.wikipedia.org/wiki/Fair_and_Accurate_Credit_Transactions_Act
' (length=71)

[Bompa]

For that, you'll have to wait for a php guy to come by.

[perkiset]

You could have more than one set of parens (captures).

preg_match returns the first array

• as what matched the entire pattern, then array[1] will be the things that matched the first set of parens, array[2] will be the second set of parens etc.

Really nice post, Bomps.

[Phaëton]

Only one set of parens... i can get the answer i need just
grab the elements from array 0 but heres the code:

but just so i can understand, what am i putting in here that makes it two arrays?

Code:

<?php
$re = '/(http:\/\/.*?\s)/';
preg_match_all( $re,
    " Vermont was the first state to pass more aggressive privacy policies 
with respect to credit reports in 1992. Maine and California are typically
 early adopters when caring for their state citizen’s privacy.
       http://www.privacyrights.org/fs/fs6a-facta.htm 
    http://uspirg.org/uspirg.asp?id2=13649
     http://en.wikipedia.org/wiki/Fair_and_Accurate_Credit_Transactions_Act

",$out, PREG_PATTERN_ORDER);
var_dump($out);
//print_r($out);
foreach ($out[0] as $k=>$v){
    echo "k::$k::v::$v::<br><hr><br>";
}
?>

Only one set of parens..

the output:

Code:

array
  0 =>
    array
      0 => string 'http://www.privacyrights.org/fs/fs6a-facta.htm ' (length=47)
      1 => string 'http://uspirg.org/uspirg.asp?id2=13649
' (length=39)
      2 => string 'http://en.wikipedia.org/wiki/Fair_and_Accurate_Credit_Transactions_Act
' (length=71)
  1 =>
    array
      0 => string 'http://www.privacyrights.org/fs/fs6a-facta.htm ' (length=47)
      1 => string 'http://uspirg.org/uspirg.asp?id2=13649
' (length=39)
      2 => string 'http://en.wikipedia.org/wiki/Fair_and_Accurate_Credit_Transactions_Act
' (length=71)

k::0::v::http://www.privacyrights.org/fs/fs6a-facta.htm ::

--------------------------------------------------------------------------------

k::1::v::http://uspirg.org/uspirg.asp?id2=13649 ::

--------------------------------------------------------------------------------

k::2::v::http://en.wikipedia.org/wiki/Fair_and_Accurate_Credit_Transactions_Act ::

--------------------------------------------------------------------------------

[perkiset]

Yes ... if you had no parens that REGEX would work the same. You do not need parens around the whole thing to capture it.

If you ditch the parens, you'll still get an array of arrays ... but array[0][0..n] will contain your answers. Consider this:

preg_match('/([A-Z0-9]{1,5})-([A-Z0-9]{1,5})/i', $input, $parts);

The array would come back with 3 entries, then sub elements that matched the items within the parenthesis.

The deal is that you're thinking from a perspective of your targeted search, not all searches: it is entirely possible that the regex above could return MANY entries. I'm going to try to type out a quick example for demonstration:

Input value:
T12142-12345, T1334-111, T55555-12345

Return array:

$arr(0)
   (0) - T12142-12345
   (1) - T1334-111
   (2) - T55555-12345
$arr(1)
   (0) - T12142
   (1) - T1334
   (2) - T55555
$arr(2)
   (0) - 12345
   (1) - 111
   (2) - 12345

Note how the first array entry is every time the entire pattern was found, the successive arrays contain the captures I defined with parenthesis.

Hope that helps

[vsloathe]

Yes this is all supremely logical. Just be logical.

[Bompa]

 http://php.net/manual/en/function.preg-match-all.php
PREG_PATTERN_ORDER

    Orders results so that $matches[0] is an array of full pattern matches, $matches[1] is an array of strings matched by the first parenthesized subpattern, and so on.

I think perk said that in his overly complex way.

$matches[0] < That's your first array output, no parentheses needed.
$matches[1] < Second array is caused by parentheses.

Ditch the parens, I think that is your answer.

Bompa

[perkiset]

 

[Phaëton]

Thanks for your responses! I appreciate your help!

I think I get it.. I did read around before I posted....

However, I'm much more of a haptic learner so now having DONE
this I can see how cool this can be... so I could use the
parentheses to strip out the http:// and other stuff like this and go with array[2]

Nice.  Bug style parsing with collection on steroids.

$re = '/(http:\/\/(.*?)\s)/';

Code:

array
  0 =>
    array
      0 => string 'http://www.privacyrights.org/fs/fs6a-facta.htm ' (length=47)
      1 => string 'http://uspirg.org/uspirg.asp?id2=13649 ' (length=39)
      2 => string 'http://en.wikipedia.org/wiki/Fair_and_Accurate_Credit_Transactions_Act
' (length=71)
  1 =>
    array
      0 => string 'http://www.privacyrights.org/fs/fs6a-facta.htm ' (length=47)
      1 => string 'http://uspirg.org/uspirg.asp?id2=13649 ' (length=39)
      2 => string 'http://en.wikipedia.org/wiki/Fair_and_Accurate_Credit_Transactions_Act
' (length=71)
  2 =>
    array
      0 => string 'www.privacyrights.org/fs/fs6a-facta.htm' (length=39)
      1 => string 'uspirg.org/uspirg.asp?id2=13649' (length=31)
      2 => string 'en.wikipedia.org/wiki/Fair_and_Accurate_Credit_Transactions_Act' (length=63)

[perkiset]

The outermost parens are redundant. You'll catch everything in arr[0] just because it matches your pattern.

The inner parens will then show up in arr[1].

Also, I'm not sure if the Regex parse is just forgiving you, I didn't think you could nest captures...

[Phaëton]

Ah. gotcha.  thats where I initially was confused with
the double arrays... so heres what i meant to post
earlier:

$re = '/(http:\/\/(.*?))\s/';

Thank you again!  That's a dyslexic nightmare...

Code:

array
  0 =>
    array
      0 => string 'http://www.privacyrights.org/fs/fs6a-facta.htm ' (length=47)
      1 => string 'http://uspirg.org/uspirg.asp?id2=13649 ' (length=39)
      2 => string 'http://en.wikipedia.org/wiki/Fair_and_Accurate_Credit_Transactions_Act
' (length=71)
  1 =>
    array
      0 => string 'http://www.privacyrights.org/fs/fs6a-facta.htm' (length=46)
      1 => string 'http://uspirg.org/uspirg.asp?id2=13649' (length=38)
      2 => string 'http://en.wikipedia.org/wiki/Fair_and_Accurate_Credit_Transactions_Act' (length=70)
  2 =>
    array
      0 => string 'www.privacyrights.org/fs/fs6a-facta.htm' (length=39)
      1 => string 'uspirg.org/uspirg.asp?id2=13649' (length=31)
      2 => string 'en.wikipedia.org/wiki/Fair_and_Accurate_Credit_Transactions_Act' (length=63)

[Phaëton]

Is there a tighter way to write this regex... the target result being
array[1] directly above this post... I guess not since its gonna
spit out array[0] no matter what, right?

[perkiset]

No meng, you're cool. Just:

$re = '/http:\/\/(.*?)\s/';

... no outside parens.