Path: csiph.com!x330-a1.tempe.blueboxinc.net!usenet.pasdenom.info!weretis.net!feeder4.news.weretis.net!news.musoftware.de!wum.musoftware.de!fu-berlin.de!uni-berlin.de!individual.net!not-for-mail
From: Sandman <mr@sandman.net>
Newsgroups: comp.lang.php
Subject: Re: preg_match() oddities and question
Date: Tue, 22 Nov 2011 13:12:39 +0100
Lines: 114
Message-ID: <mr-C40BF0.13123922112011@News.Individual.NET>
References: <mr-5B96D1.12212022112011@News.Individual.NET> <jag24m$4nj$1@softins.clara.co.uk>
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
X-Trace: individual.net 0nS/pJo9HgCXoxcQwvipkge3OdWStiAnaBczDOSjy5GmuNSvM=
X-Orig-Path: mr
Cancel-Lock: sha1:Yfir6hoNamBBsWiKlHsQ44Y7l4I=
User-Agent: MT-NewsWatcher/3.5.2 (Intel Mac OS X)
X-Face: $@,Vfa$,)%=Qa7L]y)&oZj_\EiHc}}A<Y3TvbI2&|e"bE9zc[o<ThMgB4%*L$b1YsNl!/ <lHO$>f0Bei"4a_%)"c6TQ+P/:53>;PNGuWUmkqyeN-qM65foJ[;T_(k;>]&G\T4Lhm:2 ujye2_,iUJFE;NZn>y;.|-hl7g~bIOF1qG\o<?]4mXkW*mT3]{Bn&VwP7(M0uYnGA!V!? {"y?BkBDW6e-.=I5
X-Killfiled: yttrx, gallopinginsanity.com, Mark Kent, Maverick, Nasht.n, NRen2, MuahMan, weedhopper, PC Guy, Brian, nospam@nospam.com, Oxford, Jim Lee Jr., Mocassin Joe, zara, Chance Furlong, Robert Whelan, jt2002a@hotmail.com
Xref: x330-a1.tempe.blueboxinc.net comp.lang.php:3859

In article <jag24m$4nj$1@softins.clara.co.uk>,
 tony@mountifield.org (Tony Mountifield) wrote:

> > So I have this regexp:
> > 
> > if (preg_match("/^(.*?)\s*(\d*?)\s*([A-Z,a-z,-]*?)$/", $search, $m)){
> 
> You don't need the commas in the character class, unless you want to
> match a literal comma, in which case you only need it once.

Right, thanks :)

> >     $streetname = uc_words($m[1]);
> >     $streetnumber = trim($m[2]);
> >     $streetletter = strtoupper($m[3]);
> >     $search = trim($streetname . SPACE . $streetnumber . 
> > $streetletter);
> > }
> > 
> > The desired result is taki9ng the input ($search) and split it into 
> > its parts as an address, right? $search can be, for example, "foo 
> > street 34", "longstreet 45b", "longstreet 45 b" or just "longstreet".
> 
> What about "foo street"? (i.e. with a space, but no number)

Exactly, that gets this:

Array
(
    [0] => foo street
    [1] => foo
    [2] => 
    [3] => street
)

Which is incorrect. IN fact, the last group SHOULD be defined as 
([A-Za-z]{0,1}) but that still messes it up like:

Array
(
    [0] => foo street
    [1] => foo stree
    [2] => 
    [3] => t
)

So I've tried variations for that as well.

<snip>

> And you would also get:
> 
> Array
> (
>     [0] => foo street
>     [1] => foo
>     [2] => 
>     [3] => street
> )
> 
> > As you can see, the last group "([A-Z,a-z,-]*?)" matches the entire 
> > search term since there are no digits and the first group is 
> > non-greedy. And if I make the first group greedy, "longstreet" is 
> > matched correctly, but it also catches the entire "longstreet 45b" 
> > when searching for that.
> 
> Yes, you need to define your rules more closely. Not at the regex level,
> but actually at the logic/decision level. If you can make rules that
> can unambiguously specify how all kinds of input should be parsed,
> then you can look at how to represent that in regexes. You might need
> some additional logic to operate on the parsed result.

What you're basically suggesting is a series of regexp to find out 
what "style" an adress is given in, and then parse out the parts? 
Because I'm not sure how I would be able to do it without a series if 
if/else preg_match():es?

> > Also, when searching for a term in swedish characters, I get this:
> > 
> > Array
> > (
> >     [0] => vikavÀgen
> >     [1] => vikavÀ
> >     [2] => 
> >     [3] => gen
> > )
> > 
> > Which is quite odd to me, why isn't "vikavÀgen" matched the same 
> > (undesired) way that "oongstreet". I have tried the /u modifier, and 
> > made sure that it was utf8-encoded, but it didn't make a difference 
> > (incoming encoding is ISO 8859-1).
> > 
> > Why the difference, and how do I correctly parse out parts as needed?
> 
> That's because À is not in the set A-Za-z. If you want a character class
> that properly recognises locale-specific letters, you need to change your
> character class above to this:
> 
> [[:alpha:]\-]
> 
> Hope this helps!

That explains the difference, thank you very much for that. Now I 
still need to figure out a global parse routine or criteria for 
parsing out the address parts...







-- 
Sandman[.net]