[RESOLVED] Syntax for Pattern Matching for 'preg_match_all'

sherry

Hello.

What is the syntax for the following needed for 'preg_match_all':

Objective: verify the required mandatory format is typed, else generate an error.

Simple Example: 'ZZTop' is correct but NOT 'ZZTops' nor ' ZZTop' nor 'ZZTop ' nor ' ZZTopper'

As you can see, no spaces, no added / subtracted characters, etc.

For the sake of clarity, assume its a custom bbcode and want [ZZTop=groupie] but want this format ONLY, NOTHING ELSE.

Is there a way to do that within the 'preg_match_all' statement, all in one shot, e.g. if the 'pattern' submitted doesn't match this pattern, then...

(I read somewhere how to add '^' to a character class - not sure how but I'd like to create a custom character class so I can use it anywhere, anytime...)

OR would it be e.g. a 2-step process where 1st you find the preg_match'es, then search its result array for those that don't match using some string or other function?

I've looked and haven't found the solution.

Thanks in advance.

NogDog

I think what you're looking for is the "word boundary" escape sequence:

preg_match_all('/\bZZTop\b/', $string, $matches);

sherry

Thanks for your quick reply, NogDog.

Where do I add them when the syntax includes '#' and the pattern next to 'ZZTop' is a subpattern?

I added them in various locations and no matches were found that were found without the word boundary syntax, only because I'm sure I probably didn't add them properly.

eg. '#[ZZTop(=(subpattern2)...)?]#i'

Where should I add the word boundary code in the pattern above? I want to make mandatory the subpatterns and e.g. not allow, [ZZTop] or any other permutation that's not in the correct format.

Thanks again in advance.

Weedpacket

It's not clear what the correct format is; from

sherry wrote:
As you can see, no spaces, no added / subtracted characters, etc.

Then the only string that can match "ZZTop" is .... "ZZTop".

[ZZTop=groupie]

Isn't a format, it's an example 🙂 Assuming "groupie" can be any string (except one that contains ']'):

preg_match_all('/\[ZZTop=(.*?)\]/', $string, $matches)

sherry

Thanks for your reply, Weedpacket.

The format I'm working with is:

'#[ZZTop(=(subpattern2)...)?]#i'

which I already have. It creates the output, '[ZZTop=groupie]' which is fine BUT,

it has 'HOLE' whereby allowing e.g. '[ZZTop]' to get through without an error. If it doesn't fit the correct format, I don't want the incorrect format stored in the database.

I'm trying to find a way to plug that hole, therefore, I'd like something like, 'unless this format, '[ZZTop=groupie]', then...'

I haven't found a way to use the word boundary correctly within the given
'#[ZZTop(=(subpattern2)...)?]#i'

such that any match is found. Somehow, where I put it is not correct. I need the correct syntax with #[ZZTop(=(subpattern2)...)?]#i'.

Thanks again and thanks in advance.

NogDog

Ignoring regexes and implementation details for now, can you just explain to me functionally what you want to happen? Perhaps you could show us a couple or three sample text strings with valid values and invalid values (pointing out to us slow learners which is which 😉 ). Then we can suggest a solution to achieve that functional requirement, whether it's a regex or something else.

sherry

Thanks for your reply.

If looking e.g. at vB's

tag and it's regex which has the format, '#[ZZTop(=(subpattern2)...)?]#i'.

This regex is meant to create the format, '[ZZTop=groupie]' which I like. The problem with this for my case, using this for custom tags, etc. is that it allows '[ZZTop]' which I don't want. Therefore for me, it represents a 'hole' although for others it may not be. I'm just trying to prevent [ZZTop] from being saved to database.

HTH

Thanks again and in advance.

bradgrafelman

Well, Weedpacket's pattern would allow [ZZTop=] but if you want something to be required after the equal sign as well, you would change the ".*?" to a ".+?" (from "zero or more" to "one or more").

Weedpacket

Also

'#\[ZZTop(=(subpattern2)...)?\]#i'

Sinc eyou have the =(subpattern) sign inside a ()? part, you're stating that it is optional. You'll note that in what I wrote, the = was outside, and that the ? immediately followed the *.

sherry

Thanks folks!

I'll give it a try and let you know tomorrow. I'm just signing on to check replies but have to sign off now.

Thanks again.

ttyl

P.S. This forum is soooo... COOL!!!!

sherry

Hello folks.

Thanks for ALL of your help!

After much digesting and studying regular expressions to learn it based on the feedback I received from this thread, I revisited the existing regex I was trying to use and in conjunction with the WONDERFUL info you've provided, here is the conclusion I've reached thus far:

...to use Weedpacket's '/[ZZTop=(.?)]/' with a twist of bradgrafelman to exchange + for and NogDog's intoduction to word boundary code and his suggestion [for me] to explain into words and output example as to what I want the preg_match_all or other to do.

I therefore took the time necessary to really understand alot more, not all though, and I believe I now know more on how to apply regexs, and for that I am TRULY GRATEFUL!

I'm sure that as I encounter more complex situations, I may have a question or two for this forum. 🙂

But for now, I have just one last (?) question...:

To prevent spaces, why isn't the code:

'/[ZZTop=(.+?\S)]/' or '/[ZZTop=(\S.+?)]/'

and how to prevent any individual character, e.g. maybe a ' or $ or ...?

Thanks again and thanks in advance.

NogDog

If the characters you want to disallow are fewer than those you want to allow, then you could use a negated character class ("^" right after the opening "[" of the character class):

'/\[ZZTop=([^\]\s$\']+)\]/'

Alternatively, if you only want to allow letters, numbers, and/or underscores; then there's a handy escape character for that: "\w" ("word" characters):

'/\[ZZTop=(\w+)\]/'

sherry

Wow NogDog, this is great info!

Based on your example, it appears the negation of a space(s) is if it occurs at the end of the word. How to prevent having a space elsewhere, e.g. if a typo of 'grou pie'? (without the quotation marks - used here for emphasis).

Also, how/when to use '\S', then? Is this the flipside of the case you mentioned to... 'If the characters you want to disallow are fewer than those you want to allow'?

Thanks again and in advance.

Weedpacket

'/\[ZZTop=(\w+)\]/'

doesn't allow spaces anywhere in the string (only an explicit [ZZTop= followed by one or more wordlike characters (which doesn't include the space character) followed by an explicit ].

\sis any space-like character (space, tab, line break, carriage return); \S is any other character, just as \w is wordlike, \W is non-wordlike, \d is digit-like, and \D is non-digit like. So, "when to use '\S'"? When you want to match a non-whitespace character.

'/[ZZTop=(.+?\S)]/' or '/[ZZTop=(\S.+?)]/'

Because . matches the space character: b[/b] basically matches any string and only requires it to end with a non-space character. The other case is similar, but replacing "end" with "begin".

The page in the manual covering regular expressions has a lot it needs to get through, so to be feasibly short it leaves out quite a few whys and wherefores: it needs to be augmented with outside study.

sherry

I think I got it!

Thanks, Weedpacket. Yours and NogDog's explanation are so much more clear than the references (both book & Internet) I've read. Maybe you need to publish your own. I'm sure it would be a 'Best Seller!'

Thanks so much. I'm sure others will benefit from this thread. You folks are GOOD!!!

Take Care!

Edit: Of course, bradgrafelman, that includes you, too!!!