©
Dieses Dokument verwendet PHP-Handbuch für chinesische Websites Freigeben
(PHP 5 >= 5.2.0, PHP 7)
filter_var — 使用特定的过滤器过滤一个变量
$variable
[, int $filter
= FILTER_DEFAULT
[, mixed $options
]] )
variable
待过滤的变量
filter
The ID of the filter to apply. The Types of filters manual page lists the available filters.
If omitted, FILTER_DEFAULT
will be used, which is
equivalent to
FILTER_UNSAFE_RAW
.
This will result in no filtering taking place by default.
options
一个选项的关联数组,或者按位区分的标示。如果过滤器接受选项,可以通过数组的 "flags" 位去提供这些标示。 对于回调型的过滤器,应该传入 callable 。这个回调函数必须接受一个参数,即待过滤的值,并且 返回一个在过滤/净化后的值。
<?php
// for filters that accept options, use this format
$options = array(
'options' => array(
'default' => 3 , // value to return if the filter fails
// other options here
'min_range' => 0
),
'flags' => FILTER_FLAG_ALLOW_OCTAL ,
);
$var = filter_var ( '0755' , FILTER_VALIDATE_INT , $options );
// for filter that only accept flags, you can pass them directly
$var = filter_var ( 'oops' , FILTER_VALIDATE_BOOLEAN , FILTER_NULL_ON_FAILURE );
// for filter that only accept flags, you can also pass as an array
$var = filter_var ( 'oops' , FILTER_VALIDATE_BOOLEAN ,
array( 'flags' => FILTER_NULL_ON_FAILURE ));
// callback validate filter
function foo ( $value )
{
// Expected format: Surname, GivenNames
if ( strpos ( $value , ", " ) === false ) return false ;
list( $surname , $givennames ) = explode ( ", " , $value , 2 );
$empty = (empty( $surname ) || empty( $givennames ));
$notstrings = (! is_string ( $surname ) || ! is_string ( $givennames ));
if ( $empty || $notstrings ) {
return false ;
} else {
return $value ;
}
}
$var = filter_var ( 'Doe, Jane Sue' , FILTER_CALLBACK , array( 'options' => 'foo' ));
?>
Returns the filtered data, or FALSE
if the filter fails.
Example #1 一个 filter_var() 的例子
<?php
var_dump ( filter_var ( 'bob@example.com' , FILTER_VALIDATE_EMAIL ));
var_dump ( filter_var ( 'http://example.com' , FILTER_VALIDATE_URL , FILTER_FLAG_PATH_REQUIRED ));
?>
以上例程会输出:
string(15) "bob@example.com" bool(false)
[#1] yoanlin93 at gmail dot com [2015-11-22 06:34:02]
Some boolean conversions:
<?php
var_dump(filter_var('oops', FILTER_VALIDATE_BOOLEAN, array('flags' => FILTER_NULL_ON_FAILURE)));
// NULL
var_dump(filter_var('false', FILTER_VALIDATE_BOOLEAN, array('flags' => FILTER_NULL_ON_FAILURE)));
// bool(false)
var_dump(filter_var('true', FILTER_VALIDATE_BOOLEAN, array('flags' => FILTER_NULL_ON_FAILURE)));
// bool(true)
var_dump(filter_var(0, FILTER_VALIDATE_BOOLEAN, array('flags' => FILTER_NULL_ON_FAILURE)));
// bool(false)
var_dump(filter_var(1, FILTER_VALIDATE_BOOLEAN, array('flags' => FILTER_NULL_ON_FAILURE)));
// bool(true)
var_dump(filter_var('TRUE', FILTER_VALIDATE_BOOLEAN, array('flags' => FILTER_NULL_ON_FAILURE)));
// bool(true)
var_dump(filter_var('', FILTER_VALIDATE_BOOLEAN, array('flags' => FILTER_NULL_ON_FAILURE)));
// bool(false)
var_dump(filter_var('FALSE', FILTER_VALIDATE_BOOLEAN, array('flags' => FILTER_NULL_ON_FAILURE)));
// bool(false)
[#2] Anonymous [2015-11-04 11:20:30]
FILTER_VALIDATE_URL allows:
filter_var('javascript://comment%0Aalert(1)', FILTER_VALIDATE_URL);
Where the %0A (URL encoded newline), in certain contexts, will split the comment from the JS code.
This can result in an XSS vulnerability.
[#3] joe at bloe dot com [2014-07-13 15:04:49]
"(comment)localpart@example.com"
is an invalid E-Mail address per RFC5322 (Appendix A.6.3):
"Also, the comments and white space throughout addresses, dates, and message identifiers are all part of the obsolete syntax."
[#4] george at NOSPAM dot crownvalleysoftware dot com [2014-04-30 22:08:16]
In the options, "logical disjunction" means "or" - for example:
filter_var($somestring ,FILTER_SANITIZE_STRING, FILTER_FLAG_STRIP_LOW | FILTER_FLAG_STRIP_HIGH);
[#5] Tom [2013-11-18 10:23:23]
It is important to note that though the data type of the first parameter of the function is stated as "mixed", this is only one half of the truth.
While it accepts any data type, the first parameter will always be cast to string before being validated or sanitized.
It seems that this function was designed strictly to be used on user input strings. For example: from an online-form. When using it for anything other than that, you may see issues. So read the documentation very carefully!
Especially note that there is an (to date) unresolved issue (#49510) concerning the Boolean filter while using the FILTER_NULL_ON_FAILURE flag. Note that both (string) FALSE and FALSE are not recognized as boolean values and will return NULL (not FALSE as you might expect).
I thus personally suggest that (to date) the best way to take the filter_var()-functions beyond their original purpose (and allow future extension and customization) is to wrap them in your own classes. This will allow you to work-around unexpected behavior on non-string input and add your custom checks, or back-port filters or sanitizers that may be added in later versions of PHP.
(Especially since PHP currently still lacks filters and sanitizers for some of the more exotic HTML5 input types, like "color". Thus there actually is a chance that we may see a need for custom filters or backports at some point in the future.)
[#6] CertaiN [2013-08-01 10:35:10]
E-mail validator function that supports IPv6 and checking length.
This also supports Japanese old feature phone E-mail address by sending second argument $strict as FALSE.
<?php
function validate_email($email, $strict = true) {
$dot_string = $strict ?
'(?:[A-Za-z0-9!#$%&*+=?^_`{|}~\'\\/-]|(?<!\\.|\\A)\\.(?!\\.|@))' :
'(?:[A-Za-z0-9!#$%&*+=?^_`{|}~\'\\/.-])'
;
$quoted_string = '(?:\\\\\\\\|\\\\"|\\\\?[A-Za-z0-9!#$%&*+=?^_`{|}~()<>[\\]:;@,. \'\\/-])';
$ipv4_part = '(?:[0-9]|[1-9][0-9]|1[0-9][0-9]|2[0-4][0-9]|25[0-5])';
$ipv6_part = '(?:[A-fa-f0-9]{1,4})';
$fqdn_part = '(?:[A-Za-z](?:[A-Za-z0-9-]{0,61}?[A-Za-z0-9])?)';
$ipv4 = "(?:(?:{$ipv4_part}\\.){3}{$ipv4_part})";
$ipv6 = '(?:' .
"(?:(?:{$ipv6_part}:){7}(?:{$ipv6_part}|:))" . '|' .
"(?:(?:{$ipv6_part}:){6}(?::{$ipv6_part}|:{$ipv4}|:))" . '|' .
"(?:(?:{$ipv6_part}:){5}(?:(?::{$ipv6_part}){1,2}|:{$ipv4}|:))" . '|' .
"(?:(?:{$ipv6_part}:){4}(?:(?::{$ipv6_part}){1,3}|(?::{$ipv6_part})?:{$ipv4}|:))" . '|' .
"(?:(?:{$ipv6_part}:){3}(?:(?::{$ipv6_part}){1,4}|(?::{$ipv6_part}){0,2}:{$ipv4}|:))" . '|' .
"(?:(?:{$ipv6_part}:){2}(?:(?::{$ipv6_part}){1,5}|(?::{$ipv6_part}){0,3}:{$ipv4}|:))" . '|' .
"(?:(?:{$ipv6_part}:){1}(?:(?::{$ipv6_part}){1,6}|(?::{$ipv6_part}){0,4}:{$ipv4}|:))" . '|' .
"(?::(?:(?::{$ipv6_part}){1,7}|(?::{$ipv6_part}){0,5}:{$ipv4}|:))" .
')';
$fqdn = "(?:(?:{$fqdn_part}\\.)+?{$fqdn_part})";
$local = "({$dot_string}++|(\"){$quoted_string}++\")";
$domain = "({$fqdn}|\\[{$ipv4}]|\\[{$ipv6}]|\\[{$fqdn}])";
$pattern = "/\\A{$local}@{$domain}\\z/";
return preg_match($pattern, $email, $matches) &&
(
!empty($matches[2]) && !isset($matches[1][66]) && !isset($matches[0][256]) ||
!isset($matches[1][64]) && !isset($matches[0][254])
)
;
}
?>
[#7] gt at kani dot hu [2013-06-21 14:31:44]
I found some addresses that FILTER_VALIDATE_EMAIL rejects, but RFC5321 permits:
<?php
foreach (array(
'localpart.ending.with.dot.@example.com',
'(comment)localpart@example.com',
'"this is v@lid!"@example.com',
'"much.more unusual"@example.com',
'postbox@com',
'admin@mailserver1',
'"()<>[]:,;@\\"\\\\!#$%&\'*+-/=?^_`{}| ~.a"@example.org',
'" "@example.org',
) as $address) {
echo "<p>$address is <b>".(filter_var($address, FILTER_VALIDATE_EMAIL) ? '' : 'not')." valid</b></p>";
}
?>
Results:
localpart.ending.with.dot.@example.com is not valid
(comment)localpart@example.com is not valid
"this is v@lid!"@example.com is not valid
"much.more unusual"@example.com is not valid
postbox@com is not valid
admin@mailserver1 is not valid
"()<>[]:,;@\"\\!#$%&'*+-/=?^_`{}| ~.a"@example.org is not valid
" "@example.org is not valid
The documentation does not saying that FILTER_VALIDATE_EMAIL should pass the RFC5321, however you can meet with these examples (especially with the first one). So this is a note, not a bug report.
[#8] drew_mirage at hotmail dot com [2013-06-18 19:14:16]
One key thing to remember about filtering integers is that the value for the option max_range must be less than or equal to the value of PHP_INT_MAX.
filter_var($someVariable, FILTER_VALIDATE_INT, array('options' => array('min_range' => 1, 'max_range' => SOME_VALUE_GREATER_THAN_PHP_INT_MAX)));
This will fail even if $someVariable is a valid integer in the expected range.
This can show up when you are attempting to validate a potential key for an unsigned MySQL INT type (whose maximum value is 4294967295) on a 32-bit system, where the value of PHP_INT_MAX is 2147483647.
[#9] cabrinosimone at gmail dot com [2013-04-02 15:19:35]
Pay attention that the function will not validate "not latin" domains.
if (filter_var('???????@???.???', FILTER_VALIDATE_EMAIL)) {
echo 'VALID';
} else {
echo 'NOT VALID';
}
[#10] alex4home at gmail dot com [2013-01-10 00:18:02]
Keep in mind that FILTER_VALIDATE_EMAIL will validate the email address according to standards.
However, giving the fact that organizations are free to restrict the forms of their own email addresses, using ONLY this filter can you a lot of bounces.
gmail, yahoo, hotmail, aol have special rules
For example :
<?php
$email_a = '0hot\'mail_check@hotmail.com';
if (filter_var($email_a, FILTER_VALIDATE_EMAIL)) {
echo "This (email_a) email address is considered valid.";
//reported as valid
}
//there can be no "0hotmail_check@hotmail.com"
//because hotmail will say "Your email address needs to start with a letter. Please try again." even if you remove the '
?>
[#11] marcus at synchromedia dot co dot uk [2012-09-03 09:13:50]
It's very likely that you actually want to detect all reserved ranges, not just private IPs, and there's another constant for them that should be bitwise-OR'd with it.
<?php
function is_private_ip($ip) {
return !filter_var($ip, FILTER_VALIDATE_IP, FILTER_FLAG_NO_PRIV_RANGE | FILTER_FLAG_NO_RES_RANGE);
}
?>
[#12] Martin L [2012-07-27 10:35:29]
FILTER_SANITIZE_EMAIL header injection test.
<?php
$InjString = "\r\n|\n|%0A|%0D|bcc:|to:|cc:|Content-Type:|Mime-Type:|";
echo filter_var($InjString, FILTER_SANITIZE_EMAIL);
?>
||%0A|%0D|bcc|to|cc|Content-Type|Mime-Type|
[#13] Andi, info at pragmamx dot org [2012-05-23 18:07:54]
And this is also a valid url
http://example.com/"><script>alert(document.cookie)</script>
[#14] keevitaja at gmail dot com [2012-01-28 20:05:59]
please note FILTER_VALIDATE_URL passes following url
http://example.ee/sdsf"f
[#15] joelhy [2011-07-06 02:44:26]
For those looking for private ip checking, there it is:
<?php
function is_private_ip($ip)
{
return !filter_var($ip, FILTER_VALIDATE_IP, FILTER_FLAG_NO_PRIV_RANGE);
}
?>
[#16] Luke America [2011-06-04 18:09:06]
And ... if you also want to handle pre-encoded multi-byte international URL's, you can include the additional code here:
<?php
// convert multi-byte international url's by stripping multi-byte chars
$uri = urldecode($uri) . ' ';
$len = mb_strlen($uri);
if ($len !== strlen($uri))
{
$convmap = array(0x0, 0x2FFFF, 0, 0xFFFF);
$uri = mb_decode_numericentity($uri, $convmap, 'UTF-8');
}
$uri = trim($uri);
// now, process pre-encoded MBI's
$regex = '#&([a-z]{1,2})(?:acute|cedil|circ|grave|lig|orn|ring|slash|th|tilde|uml);#i';
$uri_test = preg_replace($regex, '$1', htmlentities($uri, ENT_QUOTES, 'UTF-8'));
if ($uri_test != '') {$uri = $uri_test;}
?>
[#17] php at maisqi dot com [2011-05-27 07:11:15]
FILTER_VALIDATE_URL does not support internationalized domain name (IDN). Valid or not, no domain name with Unicode chars on it will pass validation.
We can circumvent this with a home grown solutions, but C code is C code, so I've gone for the code bellow, which builds on filter_var().
<?php
$res = filter_var ($uri, FILTER_VALIDATE_URL);
if ($res) return $res;
// Check if it has unicode chars.
$l = mb_strlen ($uri);
if ($l !== strlen ($uri)) {
// Replace wide chars by ??X??.
$s = str_repeat (' ', $l);
for ($i = 0; $i < $l; ++$i) {
$ch = mb_substr ($uri, $i, 1);
$s [$i] = strlen ($ch) > 1 ? 'X' : $ch;
}
// Re-check now.
$res = filter_var ($s, FILTER_VALIDATE_URL);
if ($res) { $uri = $res; return 1; }
}
?>
The logic is simple. A non-ascii char is more than one byte long. We replace every one of those chars by "X" and check again.
An alternative will be to punycode the URI before calling filter_var(), but PHP lacks native support for punycode. I think my approach is effective. Please e-mail me if you think otherwise or see room for improvement.
[#18] drtebi at yahoo [2009-11-24 23:55:09]
Notice that filter_var with FILTER_VALIDATE_EMAIL does not work if you are trying to get a String from an XML document e.g. via xpath.
I often use XML files as configuration files and use a function that returns a string from the config file via xpath. While this worked fine before 5.2.11, it doesn't anymore (and shouldn't, since it's an XML Element, not a String).
To overcome this problem, $variable can be type-casted:
<?php
$variable = fancyXmlGetFunction('from');
filter_var((String) $variable, FILTER_VALIDATE_EMAIL);
?>
[#19] tedivm at tedivm dot com [2009-06-22 13:28:10]
How to pass options and flags-
<?php
$options = array();
$options['options']['min_range'] = 1;
$options['options']['max_range'] = 10;
$options['flags'] = FILTER_FLAG_ALLOW_OCTAL;
filter_var(3, FILTER_VALIDATE_INT, $options);
?>
[#20] jon dot bertsch at ucop dot edu [2009-03-24 07:49:26]
Here's an actual example of the filter syntax with a flag since there doesn't appear to be a one liner for this anywhere:
'hours' => array('filter'=>FILTER_SANITIZE_NUMBER_FLOAT, 'flags' => FILTER_FLAG_ALLOW_FRACTION, 'options'=> '.')
[#21] dyer85 at gmail dot com [2008-11-03 02:00:00]
Note that when using FILTER_VALIDATE_INT along with the FILTER_FLAG_ALLOW_HEX flag, the string "2f", for example, is not validated successfully, because you must use the "0x" prefix, otherwise, it treats the data as base 10.
The range options are also smart enough to recognize when the boundaries are exceeded in different bases.
Here's an example:
<?php
$foo = '256';
$bar = '0x100';
var_dump(validate_int($foo)); // false, too large
var_dump(validate_int($bar)); // false, too large
function validate_int($input)
{
return filter_var(
$input,
FILTER_VALIDATE_INT,
// We must pass an associative array
// to include the range check options.
array(
'flags' => FILTER_FLAG_ALLOW_HEX,
'options' => array('min_range' => 1, 'max_range' => 0xff)
)
);
}
?>
[#22] visseraj at gmail dot com [2008-08-28 10:31:34]
Here are the other possible flags that you can use:
http://us3.php.net/manual/hu/ref.filter.php
[#23] dale dot liszka at gmail dot com [2008-07-09 10:15:12]
Here is how to use multiple flags (for those who learn better by example, like me):
<?php
echo "|asdf".chr(9).chr(128)."_123|";
echo "\n";
// "bitwise conjunction" means logic OR / bitwise |
echo filter_var("|asdf".chr(9).chr(128)."_123\n|" ,FILTER_SANITIZE_STRING, FILTER_FLAG_STRIP_LOW | FILTER_FLAG_STRIP_HIGH);
?>
[#24] dale dot liszka at gmail dot com [2008-07-09 09:54:17]
Using the FILTER_CALLBACK requires an array to be passed as the options:
<?php
function toDash($x){
return str_replace("_","-",$x);
}
echo filter_var("asdf_123",FILTER_CALLBACK,array("options"=>"toDash"));
// returns 'asdf-123'
?>
[#25] John [2007-07-26 12:35:45]
I managed to get this to work with PHP 5.1.6 on CentOS 5 with minor difficulty.
1) Download the PECL filter package
2) Extract the tarball
3) phpize the directory
4) ./configure
5) make
6) filter-0.11.0/logical_filters.c:25:31: error: ext/pcre/php_pcre.h: No such file or directory
7) find / -name php_pcre.h
8) Make sure php-devel is installed
9) Edit filter-0.11.0/logical_filters.c and replace "ext/pcre/php_pcre.h" with the absolute path of php_pcre.h
10) make
11) make install
12) add "extension=filter.so" to php.ini
13) Restart Apache