preg_match help

[zslash]

so, have abit of a problem. seems tho my pregmatch is done wrong somehow i cant understand... lol anyways, let me try to explain.

preg_match('/[^A-Za-z ]/',$name) //i only want to allow a-z and spaces.

now heres the part i dont understand.
$name = "test test" <-- returns false (works)
$name = "test test2" <-- returns true (works)
$name = "%" <-- returns false (does not work. should not allow this)
$name = "test %¤#" <-- returns false (also should not allow this)

even tho i have ONLY a-z and spaces allowed, it still says theres nothing else even tho there is... help?

(also i need to run a check because i want real time error messages to the browser to indicate this $name is not allowed, i dont just want to remove it)

[bradgrafelman]

Works as expected for me; test code:

$testCases = array( 'test test', 'test test2', '%', 'test %¤#' );

foreach( $testCases as $test )
    {
    echo str_pad( "$test", max( array_map( 'strlen', $testCases ) ) ) . " | ";
    var_dump( preg_match('/[^A-Za-z ]/', $test ) );
    }

output:

test test  | int(0)
test test2 | int(1)
%          | int(1)
test %ñ#   | int(1)

[zslash]

well actually, it was my bad lol. i had already filtered the text i wanted to check ontop of the script... what a noob i am lol xD

[traq]

zslash;11027867 wrote:

well actually, it was my bad lol. i had already filtered the text i wanted to check ontop of the script... what a noob i am lol xD

Well, notwithstanding:

preg_match('/[^A-Za-z ]/',$name) //i only want to allow a-z and spaces.
                                 //ur doing it wrong

bradgrafelman;11027835 wrote:

Works as expected for me...

I think what bradgrafelman is pointing out is that your code is producing the correct output (i.e., "as expected" considering what you wrote). However, your comment indicates that you actually want something else: only a-z and spaces.

If I misread that, let me know.

[font=monospace][^A-Za-z ][/font] means "match if any characters in the string are not a-z or a space" - that's why "test test" -the only test string that really is "a-z and spaces"- is the only one that fails.

"match if all characters are a-z (case-insensitive) or spaces" looks like this:

[B]/^[ A-Za-z]*$/[/B]

^         [COLOR="#A9A9A9"]// start-of-line anchor (i.e., match from the beginning)[/COLOR]
[         [COLOR="#A9A9A9"]// begin character class[/COLOR]
          [COLOR="#A9A9A9"]// (<-- that's a space over there)[/COLOR]
  A-Z     [COLOR="#A9A9A9"]// uppercase "A" through "Z"[/COLOR]
  a-z     [COLOR="#A9A9A9"]// lowercase "a" through "z"[/COLOR]
]         [COLOR="#A9A9A9"]// end character class[/COLOR]
*         [COLOR="#A9A9A9"]// zero or more of those characters[/COLOR] 
$         [COLOR="#A9A9A9"]// end-of-line anchor (i.e., match until the end)

// using ^ and $ means you match against the entire string.
// therefore, you're looking for *only* those characters, instead of those characters *and/or anything else*.

// (or, use pattern modifiers)[/COLOR]
/^[ a-z]*$/i  [COLOR="#A9A9A9"]// "i" means "case-insensitive."[/COLOR]

[bradgrafelman]

@: The problem was that the strings being tested weren't as posted in the OP's first post, hence why I got different results. Also, there's no need to change the regexp pattern - only how it is used; testing if all of the characters are valid yields the same result as testing if any of them are invalid.

[traq]

ahh... I overlooked how that was being used (sorry!).