[Resolved] Pattern Matching - finds no match - WHY?

sherry

Hello.

I am working on a form and I am trying to use pattern matching.

This one works fine, whereby it only allows what is in the ereg() pattern:

if (ereg("([^{a-zA-z0-9.\,'\&} -])", $company, $arr3)) {$clean_up[company] = $arr3[0];}

This one doesn't find the characters in the ereg(): 

if (ereg("([\{\}\[\]\<\>])", $field2, $arr6)) {$clean_up[field2] = $arr6[0];} 

I tried it out on this test data: [email]afda.gsgs-hdhgf.dfg> and the '>' wasn't detected. There was no error message, either. The '>' was ignored. I even tried using the 'OR' pipe in between each one.

Why doesn't it work?  I would like it to:
 (1) find at least 1 of these characters regardless of its location within the string;

(2) find all instances of any of these within the string.

Please help me! 

Thanks. 
Sherry

LordShryku

Well, before you bang your head on this, what do you plan to do with them once you find them? Strip them out? Replace them with something else? Pattern matching is really not needed for these single characters. You could use just simple string functions like [man]strpos[/man] or [man]str_replace[/man].

As for finding all instances, it again really depends what you want to do. If you just want to find them and do nothing with them, [man]preg_match_all[/man] would work. If you wanted to replace them, you could just str_replace, since it handles all matches in the string.

sherry

Thank you for your reply.

I'm using $PHP_SELF and displaying the message re: what is or isn't allowed in a particular field and therefore providing that person an opportunity to revise their entries, especially if it was a typo. I'm doing it this way b/c I want the data to be as 'clean' and accurate as possible before it is stored in the database or data file.

I'll give the preg_match or strpos a try, but why doesn't ereg() work in the 2nd case but the 1st case does? I am totally perplexed. In the 2nd case, I am only trying to find the 1st instance of a match of any of the characters listed in order to trigger the 'clean' routine. I am not really cleaning the field; that routine just contains the appropriate error message that is associated with that particular field.

Please reply.
Thank you.

LordShryku

Using an or operator worked for me.
ereg("[<|>|]|[|{|}]", $string);

Considering the characters you're using, you may want to just use [man]htmlentities[/man] and store the characters in their html entity formats instead...

sherry

Thank you so much. That worked fine. In addition to you adding the 'or' pipe, the difference between yours and mine is that I 'escaped' '<' and '>' and the positions of the braces and brackets are different. For some reason, it seemed that ereg() didn't like having ] next to ], either. Changing it with the braces the way you had it worked.

(edited) NOTE: I just tested ']' at the end of the field text e.g. 'eee.ooo.lll]', and it was ignored when using ereg("[<|>|]|[|{|}]", $string), but its counterpart '[' worked fine. Why is ereg() inconsistent, here?

Also, thank you for pointing out the 'preg_match_all' option and reminding me about 'strpos().' According to the write-up on strpos(), it uses less overhead than ereg(), which of course can be a good thing.

You have been extremely helpful. It is greatly appreciated!

Take care.
Sherry

sherry

refer to post http://www.phpbuilder.com/board/showthread.php?threadid=10279919 for the resolution of these ignored char ']'.