[Resolved] Validating input

[Calahan]

I know where the problem lies in my code. It's with this:

if (!preg_match('/^[A-z\\'\\.\\-]$/',$value))

I've read all the notes I can find on preg_match and a really good tutorial for checking e-mail addresses but it hasn't helped me here.
What I want to check is to make sure the string contains nothing other than letters, an apostrophe, a period, or a hypen.
Can someone point out what I'm not doing correctly. Please don't write out the correct syntax, just a hint of where I'm going wrong would be good. I have to learn preg_match and it will stick in my head if I do it myself (well thats the plan 🙂)

Cheers.

[bubblenut]

Well this will fail on anything more than one character long. I'm not sure how to word this without giving the answer away straight. when you do a character class ( [A-z\'.-] ) you also have to state how many of that class you want to allow (none or more, 1 or more, or a set value/range) otherwise it will default to just one character. Just one point which I'm not sure is a problem. In your post you mentioned the apostrophe, period and hyphen in the singular, does this mean you can have only one of each in the string?
HTH
Bubble

[swr]

I suspect A-z won't do what you want. Try a-zA-Z instead.

When you have a character class (like you do in your regexp, with the square brackets [ ]) it only matches a single character. You'll need to add something after that to say "match this many characters"...

[a-z] matches a single letter
[a-z]* matches zero or more letters
[a-z]+ matches one or more letters
[a-z]{5,10} matches any word with 5 - 10 letters

You probably want this:

if (!preg_match('/^{[a-zA-Z'.-]*$/',$value))}

[bubblenut]

A-z will work but you'll get a couple of extra character in there (doesn't usually matter too much) have a look here and you'll see that if you go from A to z you'll include []^_` in the class as well
HTH
Bubble

[Calahan]

I hate it when I don't get it!

Ok, that kinda makes sense. So in this example which I use use validate e-mails:

if(!preg_match('/^[A-z]+\.[A-z][e-mail]$/',$value))

the + sign is actually saying 1 or more characters? I thought it was saying "Accept anything between A-z and a period and anything after that period within A-z and it must contain @work.org.

Is this not the case then?

Also,

your posted solution reads:if (!preg_match('/^{[a-zA-Z'.-]*$/',$value))}

Why do I not have to escape the apostrophe?
Would if (preg_match('/^{[a-zA-Z.\'-]*$/',$value))}
work too?

[bubblenut]

You do have to escape the apostrophe but vb sometimes strips the backslashes.

the + sign is actually saying 1 or more characters? I thought it was saying "Accept anything between A-z and a period and anything after that period within A-z and it must contain @work.org.

/a+/ will match "a" it will also match "aaaa" and it will match "aaaaaaaaaaaa" but it will not match "" or "ab" or "bb". The + sign says we must have 1 or more of the character before me, if that is a character class or a grouping ( something encapsulated in parenthesis) then it act on that.

With regards to writing a regex for acception an email address, this can get very complex indeed. I have a book with an email regex 6,598 bytes long!! However, it is not often you need a regex this strict, you stated your regex need only accept email addresses from the domain work.org and you quoted the regex
/^{[A-z]+.[A-z][e-mail]$/}
This is fine if you know that the email address will not have any numbers or underscores and that it will have only one period in it (this is infact the case with company emails where i work) however if this cannot be guarunted then we need to little more refining.
Bubble

[swr]

The difference between the two is that using will not require a letter before/after the dot. With a it will match [e-mail] or [e-mail] or [e-mail]; using + instead won't. This is because the * means ZERO or more of the specified character(s), while + means at least one.

I still think you should use A-Za-z instead of A-z. Using A-z will probably allow things like bl[.ah]@work.org, because (if I recall correctly) the square brackets are in that range.

Also, you'll need to escape (put a backslash before) the dots. Dot is a special character that means "any character". It's as if you had [inserteverypossiblecharacterhere] in your regular expression. This is useful (.* will match any number of any character) but what you want is a literal dot.

So it sounds like this is what you want:

/^{[A-Za-z]+.[A-Za-z][e-mail]$/}

[Calahan]

I removed the last post while you were writing. Sorry! I just figured it out and didn't want to look any more stupid than I currently feel. 😉

Your solution was exactly what I finally came up with.

Thank you both for your help. I think I can feel it starting to sink in...