foreach ($tests as $element) { if (is_numeric($element)) { echo var_export($element, true) . " is numeric", PHP_EOL; } else { echo var_export($element, true) . " is NOT numeric", PHP_EOL; } } ?>
The above example will output:
'42' is numeric
1337 is numeric
1337 is numeric
1337 is numeric
1337 is numeric
1337.0 is numeric
'0x539' is NOT numeric
'02471' is numeric
'0b10100111001' is NOT numeric
'1337e0' is numeric
'not numeric' is NOT numeric
array (
) is NOT numeric
9.1 is numeric
NULL is NOT numeric
'' is NOT numeric
Example #2 is_numeric() with whitespace
<?php $tests = [ " 42", "42 ", "\u{A0}9001", // non-breaking space "9001\u{A0}", // non-breaking space ];
foreach ($tests as $element) { if (is_numeric($element)) { echo var_export($element, true) . " is numeric", PHP_EOL; } else { echo var_export($element, true) . " is NOT numeric", PHP_EOL; } } ?>
Output of the above example in PHP 8:
' 42' is numeric
'42 ' is numeric
' 9001' is NOT numeric
'9001 ' is NOT numeric
Output of the above example in PHP 7:
' 42' is numeric
'42 ' is NOT numeric
' 9001' is NOT numeric
'9001 ' is NOT numeric
So this function is not intimidated by super-big numbers. I hope this helps someone.
PS: Also note that if you write is_numeric (45thg), this will generate a parse error (since the parameter is not enclosed between apostrophes or double quotes). Keep this in mind when you use this function.
is_numeric fails on the hex values greater than LONG_MAX, so having a large hex value parsed through is_numeric would result in FALSE being returned even though the value is a valid hex number
Note that this function is not appropriate to check if "is_numeric" for very long strings. In fact, everything passed to this function is converted to long and then to a double. Anything greater than approximately 1.8e308 is too large for a double, so it becomes infinity, i.e. FALSE. What that means is that, for each string with more than 308 characters, is_numeric() will return FALSE, even if all chars are digits.
regarding the global vs. american numeral notations, it should be noted that at least in japanese, numbers aren't grouped with an extra symbol every three digits, but rather every four digits (for example 1,0000 instead of 10.000). also nadim's regexen are slightly suboptimal at one point having an unescaped '.' operator, and the whole thing could easily be combined into a single regex (speed and all).
adjustments:
<?php
$eng_or_world = preg_match
('/^[+-]?'. // start marker and sign prefix
'(((([0-9]+)|([0-9]{1,4}(,[0-9]{3,4})+)))?(\\.[0-9])?([0-9]*)|'. // american
'((([0-9]+)|([0-9]{1,4}(\\.[0-9]{3,4})+)))?(,[0-9])?([0-9]*))'. // world
'(e[0-9]+)?'. // exponent
'$/', // end marker
$str) == 1;
?>
i'm sure this still isn't optimal, but it should also cover japanese-style numerals and it fixed a couple of other issues with the other regexen. it also allows for an exponent suffix, the pre-decimal digits are optional and it enforces using either grouped or ungrouped integer parts. should be easier to trim to your liking too.