Regular Expressions

Anon

Can anyone please help with this. i have been agonising over this for hours now. I want to extract the text between the html paragraph tags and insert them into an array. Ereg is the right function for this except I can't work out how to use Regular Expressions.

The paragraph tags contain a lot of formatting information, hence the attempt below. If anyone knows a better way of doing this I would love to know.

Here is the Reg Exp:
eregi("^{<p[a-z0-9\=-:;\']>[a-z0-9;&]$",$fileContents,$dbContents);}

and this is a typical example of what I am trying to extract:

Thanks for any help in advance.

L Ricards

Anon

Try:

preg_match("/<p.?>.?<\/p>/", $res, $res1);

Good luck,

Marc

Anon

Oooh, the dreaded Office2000 HTML bastardisations 🙂

My first cut of this would be:

preg_match_all('#<p(?:.?)>(.?)#i', $fileContents, $matches);
$dbContents=$matches[1];

Anon

I think for cut number 1.0.1, I'd replace "#i" with "#is", in case the paragraph contains some newline characters (as it most likely would).

mvdongen

Sorry, I was a bit too fast, check out the reply by Grumbler instead.

Anon

Thanks Grumbler and Marc, much appreciated. Not sure how it works, doesn't even look like the Reg Exp I was reading about. Still thanks, you have saved me a lot of time.

Luke Richards

Anon

Notice that I used a PCRE preg function instead of a POSIX ereg one. The syntax for the two are somewhat different, that for the former being described in the manual.

The PCRE syntax is quite a bit more flexible (and the functions run faster), though if you whack a regexp written in it over the head enough you could force it into POSIX syntax. You wouldn't want to have to, though.

Looking at what I wrote, I realised I didn't need to group the contents of the tag (in fact, I explictly say I'm not interested in it with the "?:"); I could have just said

"#<p.?>(.?)#is"

Bit by bit:

'#' starts the regexp. Traditionally, '/' is used, but I'll be wanting to use that character in the regexp, and I'm too lazy to bother with escaping it.

'<p' Obviously starts the paragraph tag; maybe '<\s*?p' would be smarter, as there may be some whitespace between the '<' and the 'p' ('\s' is PCRE-speak for "whitespace character").

".*?>"

Match zero or more characters, and then a ">". Without the ? this would happily gobble the ">" character at the end of the tag as well and just keep going, but we don't want to do that. The ? tells it to stop at the first ">" it finds.

Incidentally, if the regexp engine can't get a match if this part of the expressions stops at the first ">", it will try again by stopping it at the second ">".

"(.*?)<"

Again, match zero or more characters, this time as far as the first "<" it sees. This is the bit between the tags, which is what we're after, so I'll group it and store it.

The note I gave above is not so incidental now: The paragraph could contain other tags as well, so you'll want this part to match more and more "<" characters untill it sees one that is followed by:

"/p>"

Which is the end of the paragraph. If I'd used the traditional '/' characters to delimit the regexp, this bit would have had to read "\/p>".

"#is"

End of the regexp, followed by some flags that affect what matches what. 'i' for case-insenstive matching, 's' to indicate that "." should match newline characters as well.

preg_match_all() finds, as the name implies, all the matches of the supplied pattern in the given string, stashing them all into an array. I used $matches. $matches[0] contains an array which lists all strings found that match the entire regexp, while $matches[1] contains an array which lists all strings found that match the first "()"-grouped section.