what is greedy?

[dalek]

could someone explain to me in very simple terms ( ) what is meant by greedy matching in regexps? preferably via an example??

cheers!

[Egghead]

http://javaalmanac.com/egs/java.util.regex/Greedy.html

[Chris82]

Hi.

Greedy means that the regexp matches from the first occurrence of the pattern until the last.

example: this is the string and we want to capture the words inside the bold tags.

$string = 'This is my <b>black</b> cat her <b>name</b> is Furball.'; 

Now if you use a "greedy" regexp like:

preg_match('#<b>(.*)</b>#', $string, $match);
print_r($match); 

The bold tags surround the result. Since this regexp is "greedy" matches from the first "<b>" to the last "</b>" in the string, the result is:

black</b> cat her <b>name

Not what we want. If we use the "?" to make the regexp ungreedy we get the desired result.

preg_match('#<b>(.*?)</b>#', $string, $match);
print_r($match); 

which results in an array containing "black" and "name".

[website]

Lets take this string for example:

$string = 'Here is | an example code | of something big | and strange'; 

Then lets run this command:

$string = preg_replace('#^(.+)\|#', 'replacement', $string); 

When doing echo($string) it would output 'replacement and strange' because it replaces from the start of the string to the last occurance of |.
This is called greedy.

However if we would run this command

$string = preg_replace('#^(.+?)\|#', 'replacement', $string); 

echo($string) would output 'replacement an example code | of something big | and strange', see how it goes from the start of the string to the first occurance of |, this is called ungreedy.
To switch from greedy to greedy use ? or the 'U' modifier.

Hope this helps

Edit:

Too slow

[Chris82]

Edit:

Too slow

Ah you're not the only one. Happened to my quite often aswell

[dalek]

thanks! i *think* i get it lol!!!

i will have a look in more depth when i get home...for example if i did a replace that swapped my custom formatting tag [b ][ eb ] for html bold tags, and did a replace on a string like

[ b ]here[ eb ] are some [ b ]words[ eb ], some of them are in [ b ]bold[ eb ] and some of them are [ b ]not[ eb ]

i would get something like
< b >here[ eb ] are some [ b ]words[ eb ], some of them are in [ b ]bold[ eb ] and some of them are [ b ]not< / b >

is that right?

[Chris82]

Yes if you use a greedy regexp that will be the result.