PHP Pattern Matching - Is there a better solution?

One of my weaknesses has always been pattern matching. Something I
definitely need to study up on and maybe you guys can give me a pointer
here.

I'm looking to remove all of this code and just use pattern matching to
determine if the proper amount of numeric characters has been met. Here is
the function I've already done. Any help you can give in a pattern matching
solution would be much appreciated and very educational.


//the $area option is for determining if the user need (or need not) type in
their area code.

function check_phone($phone,$area='y'){
if($area=='y'){
$min=10;
} else {
$min=7;
}

//is there a more direct approach to just count the amount of numbers in
the string by pattern matching?
$phone=str_replace('-','',$phone);
$phone=str_replace('(','',$phone);
$phone=str_replace(')','',$phone);
$phone=str_replace(' ','',$phone);

if(!preg_match('/[\d]{'.$min.',}/',$phone)){
return(FALSE);
} else {
return(TRUE);
}
}

.oO(gsv2com)

> //is there a more direct approach to just count the amount of numbers in
>the string by pattern matching?
> $phone=str_replace('-','',$phone);
> $phone=str_replace('(','',$phone);
> $phone=str_replace(')','',$phone);
> $phone=str_replace(' ','',$phone);
>
> if(!preg_match('/[\d]{'.$min.',}/',$phone)){
> return(FALSE);
> } else {
> return(TRUE);
> }
> }
>

I would simply remove all the special chars and then check the length of
the remaining string, no need for pattern matching here:

function check_phone($phone, $area = TRUE) {
$min = $area ? 10 : 7;
return strlen(str_replace(array('-', '(', ')', ' '), '', $phone)) >= $min;
}

HTH
Micha

Regular expression is definitely something that's worth spending the
time mastering. In this case it's not strictly necessary. But for more
complex validation, it'll makes your life a whole lot easier.

Here's the pattern, built in pieces for educational purpose:

$pattern = '';

// the beginning of the string
$pattern .= '^';

// 0 or more white spaces (just in case)
$pattern .= '\s*';
if($area) {
// optional open paren
$pattern .= '\(?';

// capturing 3 digits
$pattern .= '(\d{3})';

// optional close paren
$pattern .= '\)';

// optional dash, or space
$pattern .= '[\- ]?';
}

// capturing next 3 digits
$pattern .= '(\d{3})';

// optional dash or space
$pattern .= '[\- ]?';

// capturing 4 digits
$pattern .= '(\d{4})';

// 0 or more white spaces
$pattern .= '\s*';

// end of string
$pattern .= '$';

// altogether >> '^\s*\(?(\d{3})\)?[\- ]?(\d{3})[\- ]?(\d{4})\s*$'
if(preg_match("/$pattern/", $phone, $matches)) {
...
}

gsv2com wrote:

> One of my weaknesses has always been pattern matching. Something I
> definitely need to study up on and maybe you guys can give me a pointer
> here.
>
> I'm looking to remove all of this code and just use pattern matching to
> determine if the proper amount of numeric characters has been met. Here is
> the function I've already done. Any help you can give in a pattern matching
> solution would be much appreciated and very educational.

function check_phone($phone,$area='y'){
if($area=='y'){
$min=10;
} else {
$min=7;
}
// remove any non-digit from the string
$phone=preg_replace('/[^0-9]/','',$phone);

return (strlen($phone)==$min)? TRUE : FALSE;
}

Just be careful what you pass for the $area value. If you call:

check_phone($phone,0);

You *will* need the area code. (string to int conversion in if statement).

--
Justin Koivisto - justin@koivi.com
http://www.koivi.com

chernyshevsky@hotmail.com wrote:

> Regular expression is definitely something that's worth spending the
> time mastering. In this case it's not strictly necessary. But for more
> complex validation, it'll makes your life a whole lot easier.
>
> Here's the pattern, built in pieces for educational purpose:
>
> $pattern = '';
>
> // the beginning of the string
> $pattern .= '^';
>
> // 0 or more white spaces (just in case)
> $pattern .= '\s*';
> if($area) {
> // optional open paren
> $pattern .= '\(?';
>
> // capturing 3 digits
> $pattern .= '(\d{3})';
>
> // optional close paren
> $pattern .= '\)';

That isn't optional... this is ;)
$pattern .= '\)?';

> // optional dash, or space
> $pattern .= '[\- ]?';
> }
>
> // capturing next 3 digits
> $pattern .= '(\d{3})';
>
> // optional dash or space
> $pattern .= '[\- ]?';
>
> // capturing 4 digits
> $pattern .= '(\d{4})';
>
> // 0 or more white spaces
> $pattern .= '\s*';
>
> // end of string
> $pattern .= '$';
>
> // altogether >> '^\s*\(?(\d{3})\)?[\- ]?(\d{3})[\- ]?(\d{4})\s*$'
> if(preg_match("/$pattern/", $phone, $matches)) {
> ...
> }

Hmm.. might want to make that a bit more generic. I've seen users
delimit phone number parts with characters other than a - or space (like
a dot, long hyphen, comma, etc.).

Maybe better would be something like this:

^[^\d]*(\d{3})[^\d]*(\d{3})[^\d]*(\d{4})[^\d]*$

Of course, you could just remove all the non-digits:
$phone=preg_replace('/[^\d]/','',$phone);

Then simply compare the length of the string.

--
Justin Koivisto - justin@koivi.com
http://www.koivi.com

Thanks for your help guys. As the end result, I used a combination of a few
of your tips to make the finalized function:

function check_phone($phone,$area=TRUE){
$min = $area ? 10 : 7;
$phone=preg_replace('/[^0-9]/','',$phone);
return (strlen($phone)==$min)? TRUE : FALSE;
}

Very nice. Thanks again.

.oO(gsv2com)

>function check_phone($phone,$area=TRUE){
> $min = $area ? 10 : 7;
> $phone=preg_replace('/[^0-9]/','',$phone);
> return (strlen($phone)==$min)? TRUE : FALSE;
>}

JFTR: It's not necessary to explicitly return TRUE or FALSE in this
case. It's redundant because the result of the comparision will already
be of type boolean:

return strlen($phone) == $min;

does the same without another unnecessary operation.

In the preg_replace() pattern you could also use the character class \d
(digits) instead of 0-9:

'/[^\d]/'

And another thing: Are you sure the numbers will always exactly be 7 or
10 digits long? At least here in Germany the length of phone numbers may
vary. That's why I used >= instead of == in my example and your first
code also suggested that.

Micha

Michael Fesser wrote:

> .oO(gsv2com)
>
>
>>function check_phone($phone,$area=TRUE){
>> $min = $area ? 10 : 7;
>> $phone=preg_replace('/[^0-9]/','',$phone);
>> return (strlen($phone)==$min)? TRUE : FALSE;
>>}
>
> JFTR: It's not necessary to explicitly return TRUE or FALSE in this
> case. It's redundant because the result of the comparision will already
> be of type boolean:
>
> return strlen($phone) == $min;
>
> does the same without another unnecessary operation.

I usually include the explicit returns just for readability sake. ;)

--
Justin Koivisto - justin@koivi.com
http://www.koivi.com

gsv2com wrote:
> One of my weaknesses has always been pattern matching.

<snip>

Perhaps you should try <http://www.weitz.de/regex-coach/>

--
<?php echo 'Just another PHP saint'; ?>
Email: rrjanbiah-at-Y!com Blog: http://rajeshanbiah.blogspot.com/