URL 函数
在线手册：中文 英文
PHP手册

get_headers

(PHP 5)

get_headers — 取得服务器响应一个 HTTP 请求所发送的所有标头

说明

array get_headers ( string $url [, int $format ] )

get_headers() 返回一个数组，包含有服务器响应一个 HTTP 请求所发送的标头。如果失败则返回 FALSE 并发出一条 E_WARNING 级别的错误信息。

如果将可选的 format 参数设为 1，则 get_headers() 会解析相应的信息并设定数组的键名。

Note:

自 PHP 5.1.3 起本函数使用默认的流上下文，其可以用 stream_context_get_default() 函数设定和修改。

Example #1 get_headers() 例子

<?php
$url = 'http://www.example.com';

print_r(get_headers($url));

print_r(get_headers($url, 1));
?>

以上例程的输出类似于：

Array
(
    [0] => HTTP/1.1 200 OK
    [1] => Date: Sat, 29 May 2004 12:28:13 GMT
    [2] => Server: Apache/1.3.27 (Unix)  (Red-Hat/Linux)
    [3] => Last-Modified: Wed, 08 Jan 2003 23:11:55 GMT
    [4] => ETag: "3f80f-1b6-3e1cb03b"
    [5] => Accept-Ranges: bytes
    [6] => Content-Length: 438
    [7] => Connection: close
    [8] => Content-Type: text/html
)

Array
(
    [0] => HTTP/1.1 200 OK
    [Date] => Sat, 29 May 2004 12:28:14 GMT
    [Server] => Apache/1.3.27 (Unix)  (Red-Hat/Linux)
    [Last-Modified] => Wed, 08 Jan 2003 23:11:55 GMT
    [ETag] => "3f80f-1b6-3e1cb03b"
    [Accept-Ranges] => bytes
    [Content-Length] => 438
    [Connection] => close
    [Content-Type] => text/html
)

参数

url

The target URL.

format

If the optional format parameter is set to non-zero, get_headers() parses the response and sets the array's keys.

返回值

Returns an indexed or associative array with the headers, or FALSE on failure.

更新日志

版本 说明
5.1.3 This function now uses the default stream context, which can be set/changed with the stream_context_set_default() function.

范例

Example #2 get_headers() example

<?php
$url = 'http://www.example.com';

print_r(get_headers($url));

print_r(get_headers($url, 1));
?>

以上例程的输出类似于：

Array
(
    [0] => HTTP/1.1 200 OK
    [1] => Date: Sat, 29 May 2004 12:28:13 GMT
    [2] => Server: Apache/1.3.27 (Unix)  (Red-Hat/Linux)
    [3] => Last-Modified: Wed, 08 Jan 2003 23:11:55 GMT
    [4] => ETag: "3f80f-1b6-3e1cb03b"
    [5] => Accept-Ranges: bytes
    [6] => Content-Length: 438
    [7] => Connection: close
    [8] => Content-Type: text/html
)

Array
(
    [0] => HTTP/1.1 200 OK
    [Date] => Sat, 29 May 2004 12:28:14 GMT
    [Server] => Apache/1.3.27 (Unix)  (Red-Hat/Linux)
    [Last-Modified] => Wed, 08 Jan 2003 23:11:55 GMT
    [ETag] => "3f80f-1b6-3e1cb03b"
    [Accept-Ranges] => bytes
    [Content-Length] => 438
    [Connection] => close
    [Content-Type] => text/html
)

Example #3 get_headers() using HEAD example

<?php
// By default get_headers uses a GET request to fetch the headers. If you
// want to send a HEAD request instead, you can do so using a stream context:
stream_context_set_default(
    array(
        'http' => array(
            'method' => 'HEAD'
        )
    )
);
$headers = get_headers('http://example.com');
?>

参见

URL 函数
在线手册：中文 英文
PHP手册
PHP手册 - N: 取得服务器响应一个 HTTP 请求所发送的所有标头

用户评论:

Backslider (16-Feb-2012 03:06)

It should be noted that rather than returning "false" on failure, this function (and others) return a big phat WARNING that will halt your script in its tracks if you do not have error reporting /warning turned off.

Thats just insane!  Any function that does something like fetch a URL should simply return false, without a warning, if the URL fails for whatever reason other than it is badly formatted.

damolp at hotmail dot com (04-Jul-2011 01:43)

I found that this function is the slowest in obtaining the headers of a page probably because it uses a GET request rather then a HEAD request. Over 10,000,000 trials of obtaining the headers of a page from a server i found the following (results in seconds).

cURL: Mean: 0.584127946. Sigma: 0.050581736.
fsocketopen: Mean: 0.622114251. Sigma: 0.263170424.
get_headers: Mean: 0.90375551. Sigma: 0.273823419.

cURL was the fastest with fsocketopens being the second fastest. I noticed as well that fsocketopen had some outliers where as cURL did not.

Lukas Najduk (11-Jan-2011 11:36)

Unfortunately there is still no useful output format to handle redirects.

This function will bring all non-broken headers into a usable format. Too bad it has to call the get_headers() funtion 2 times, but i dont see any other possibility right now.

<?php
function page_get_headers($crawl_uri) {
    // get the array values where to split the second get headers return
    $headers = get_headers($crawl_uri, 1);
    if (empty($headers)) {
        return array();
    }
    $splitmarks = array();
    foreach ($headers as $key=>$h) {
        if (is_numeric($key)) {
            array_push($splitmarks, $h);
        }
    }

    // get the "real" headers
    $headers_final = array();
    $i = 0;
    $headers = get_headers($crawl_uri);
    foreach ($headers as $head) {
        // if the value is one of the splitmarks, start the next header
        if (array_key_exists($i, $splitmarks) && $head === $splitmarks[$i]) {
            $i++;
        }  
        else {
            // get the headers name
            $tmp = explode(": ", $head, 2);
            // check if the header is already existing, if yes, parse the similar header to an array
            if (!empty($headers_final[$i-1][$tmp[0]])) {
                if (is_array($headers_final[$i-1][$tmp[0]])) {
                    array_push($headers_final[$i-1][$tmp[0]], $tmp[1]);
                }
                else {
                    $headers_final[$i-1][$tmp[0]] = array($headers_final[$i-1][$tmp[0]], $tmp[1]);
                }
            }
            else {
                $headers_final[$i-1][$tmp[0]] = $tmp[1];  
            }  
        }
    }
    return $headers_final;
}
?>

Weboide (26-Sep-2010 02:13)

Note that get_headers **WILL follow redirections** (HTTP redirections). New headers will be appended to the array if $format=0. If $format=1 each redundant header will be an array of multiple values, one for each redirection.

For example:

<?php
$url = 'http://google.com';
var_dump(get_headers($url,0));
/*array(18) {
  [0]=>  string(30) "HTTP/1.0 301 Moved Permanently"
  [1]=>  string(32) "Location: http://www.google.com/"
  [2]=>  string(38) "Content-Type: text/html; charset=UTF-8"
  [3]=>  string(35) "Date: Sun, 26 Sep 2010 00:59:50 GMT"
  [4]=>  string(38) "Expires: Tue, 26 Oct 2010 00:59:50 GMT"
  [5]=>  string(38) "Cache-Control: public, max-age=2592000"
....
  string(15) "HTTP/1.0 200 OK"
  [10]=>  string(35) "Date: Sun, 26 Sep 2010 00:59:51 GMT"
  [11]=>  string(11) "Expires: -1"
  [12]=>  string(33) "Cache-Control: private, max-age=0"
.....
}*/

/*===========================*/

var_dump(get_headers($url,1));
/*array(11) {
  [0]=>
  string(30) "HTTP/1.0 301 Moved Permanently"
  ["Location"]=>  string(22) "http://www.google.com/"
  ["Content-Type"]=>  array(2) {
    [0]=>    string(24) "text/html; charset=UTF-8"
    [1]=>    string(29) "text/html; charset=ISO-8859-1"
  }
  ["Date"]=>  array(2) {
    [0]=>    string(29) "Sun, 26 Sep 2010 01:03:39 GMT"
    [1]=>    string(29) "Sun, 26 Sep 2010 01:03:39 GMT"
  }
  ["Expires"]=>  array(2) {
    [0]=>    string(29) "Tue, 26 Oct 2010 01:03:39 GMT"
    [1]=>    string(2) "-1"
  }
  ["Cache-Control"]=>  array(2) {
    [0]=>    string(23) "public, max-age=2592000"
    [1]=>    string(18) "private, max-age=0"
  }
.....
}*/
?>

nick at innovaweb dot co dot uk (03-May-2010 01:28)

Seems like there are some people who are looking for only the 3-digit HTTP response code  - here is a quick and nasty solution:

<?php
function get_http_response_code($theURL) {
    $headers = get_headers($theURL);
    return substr($headers[0], 9, 3);
}
?>

How easy is that? Echo the function containing the URL you want to check the response code for, and voilà. Custom redirects, alternative for blocked is_file() or flie_exists() functions (like I seem to have on my servers) hence the cheap workaround. But hey - it works!

Pudding

php at hm2k dot org (28-Jan-2010 06:20)

<?php
/**
 * Fetches all the real headers sent by the server in response to a HTTP request without redirects
 *
 * @link      http://php.net/function.get_headers
 * @link      http://bugs.php.net/bug.php?id=50719
 */

function get_real_headers($url,$format=0,$follow_redirect=0) {
  if (!$follow_redirect) {
    //set new default options
    $opts = array('http' =>
        array('max_redirects'=>1,'ignore_errors'=>1)
    );
    stream_context_get_default($opts);
  }
  //get headers
    $headers=get_headers($url,$format);
    //restore default options
  if (isset($opts)) {
    $opts = array('http' =>
        array('max_redirects'=>20,'ignore_errors'=>0)
    );
    stream_context_get_default($opts);
  }
  //return
    return $headers;
}
?>

gabe at vtunnel dot com (22-Apr-2009 09:23)

In some cases, you don't want get_headers to follow redirects. For example, some of my servers can access a particular website, which sends a redirect header. The site it is redirected to, however, has me firewalled. I need to take the 302 redirected url, and do something to it to give me a new url that I *can* connect to.

The following will give you output similar to get_headers, except it has a timeout, and it doesn't follow redirects:

<?php
function get_headers_curl($url)
{
    $ch = curl_init();

    curl_setopt($ch, CURLOPT_URL,            $url);
    curl_setopt($ch, CURLOPT_HEADER,         true);
    curl_setopt($ch, CURLOPT_NOBODY,         true);
    curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
    curl_setopt($ch, CURLOPT_TIMEOUT,        15);

    $r = curl_exec($ch);
    $r = split("\n", $r);
    return $r;
}

If you do want to follow redirects, you can do something like this:

$go = 1;
$i = 1;

while ($go && $i < 6)
{
    $headers = get_headers_curl($url);
    $go = getNextLocation($headers);
    if ($go)
    {
        $url = modifyUrl($go);
    }
    $i++;
}

function getNextLocation($headers)
{
    $array = $headers;
    $count = count($array);

    for ($i=0; $i < $count; $i++)
    {
        if (strpos($array[$i], "ocation:"))
        {
                $url = substr($array[$i], 10);
        }
    }
    if ($url)
    {
        return $url;
    }
    else
    {
        return 0;
    }
}
?>

info at marc-gutt dot de (21-Jun-2008 01:04)

Should be the same than the original get_headers():

<?php
if (!function_exists('get_headers')) {
function get_headers($url, $format=0) {
    $headers = array();
    $url = parse_url($url);
    $host = isset($url['host']) ? $url['host'] : '';
    $port = isset($url['port']) ? $url['port'] : 80;
    $path = (isset($url['path']) ? $url['path'] : '/') . (isset($url['query']) ? '?' . $url['query'] : '');
    $fp = fsockopen($host, $port, $errno, $errstr, 3);
    if ($fp)
    {
        $hdr = "GET $path HTTP/1.1\r\n";
        $hdr .= "Host: $host \r\n";
        $hdr .= "Connection: Close\r\n\r\n";
        fwrite($fp, $hdr);
        while (!feof($fp) && $line = trim(fgets($fp, 1024)))
        {
            if ($line == "\r\n") break;
            list($key, $val) = explode(': ', $line, 2);
            if ($format)
                if ($val) $headers[$key] = $val;
                else $headers[] = $key;
            else $headers[] = $line;
        }
        fclose($fp);
        return $headers;
    }
    return false;
}
}
?>

php dot sirlancelot at spamgourmet dot com (05-Jun-2008 09:59)

I tried to replicate the native behavior as much as possible for systems that don't have the get_headers() function. Here it is:
<?php
if (!function_exists('get_headers')) {
function get_headers($Url, $Format= 0, $Depth= 0) {
    if ($Depth > 5) return;
    $Parts = parse_url($Url);
    if (!array_key_exists('path', $Parts))   $Parts['path'] = '/';
    if (!array_key_exists('port', $Parts))   $Parts['port'] = 80;
    if (!array_key_exists('scheme', $Parts)) $Parts['scheme'] = 'http';

    $Return = array();
    $fp = fsockopen($Parts['host'], $Parts['port'], $errno, $errstr, 30);
    if ($fp) {
        $Out = 'GET '.$Parts['path'].(isset($Parts['query']) ? '?'.@$Parts['query'] : '')." HTTP/1.1\r\n".
               'Host: '.$Parts['host'].($Parts['port'] != 80 ? ':'.$Parts['port'] : '')."\r\n".
               'Connection: Close'."\r\n";
        fwrite($fp, $Out."\r\n");
        $Redirect = false; $RedirectUrl = '';
        while (!feof($fp) && $InLine = fgets($fp, 1280)) {
            if ($InLine == "\r\n") break;
            $InLine = rtrim($InLine);

            list($Key, $Value) = explode(': ', $InLine, 2);
            if ($Key == $InLine) {
                if ($Format == 1)
                        $Return[$Depth] = $InLine;
                else    $Return[] = $InLine;

                if (strpos($InLine, 'Moved') > 0) $Redirect = true;
            } else {
                if ($Key == 'Location') $RedirectUrl = $Value;
                if ($Format == 1)
                        $Return[$Key] = $Value;
                else    $Return[] = $Key.': '.$Value;
            }
        }
        fclose($fp);
        if ($Redirect && !empty($RedirectUrl)) {
            $NewParts = parse_url($RedirectUrl);
            if (!array_key_exists('host', $NewParts))   $RedirectUrl = $Parts['host'].$RedirectUrl;
            if (!array_key_exists('scheme', $NewParts)) $RedirectUrl = $Parts['scheme'].'://'.$RedirectUrl;
            $RedirectHeaders = get_headers($RedirectUrl, $Format, $Depth+1);
            if ($RedirectHeaders) $Return = array_merge_recursive($Return, $RedirectHeaders);
        }
        return $Return;
    }
    return false;
}}
?>
The function will handle up to five redirects.
Enjoy!

dxtrim at yahoo dot com (17-Jan-2008 09:16)

Content-Type returns a value depending only on the extension and not the real MIME TYPE.

So, bad_file.exe renamed to good_file.doc will return application/msword

A file without extension returns a 404.

(13-Nov-2006 08:29)

I've noticed it.
Some Server will simply return the false reply header if you sent 'HEAD' request instead of 'GET'. The 'GET' request header always receiving the most actual HTTP header instead of 'HEAD' request header. But If you don't mind for a fast but risky method then 'HEAD' request is better for you.

btw ... this is get header with additional information such as User, Pass & Refferer. ...
<?php
    function get_headers_x($url,$format=0, $user='', $pass='', $referer='') {
        if (!empty($user)) {
            $authentification = base64_encode($user.':'.$pass);
            $authline = "Authorization: Basic $authentification\r\n";
        }

        if (!empty($referer)) {
            $refererline = "Referer: $referer\r\n";
        }

        $url_info=parse_url($url);
        $port = isset($url_info['port']) ? $url_info['port'] : 80;
        $fp=fsockopen($url_info['host'], $port, $errno, $errstr, 30);
        if($fp) {
            $head = "GET ".@$url_info['path']."?".@$url_info['query']." HTTP/1.0\r\n";
            if (!empty($url_info['port'])) {
                $head .= "Host: ".@$url_info['host'].":".$url_info['port']."\r\n";
            } else {
                $head .= "Host: ".@$url_info['host']."\r\n";
            }
            $head .= "Connection: Close\r\n";
            $head .= "Accept: */*\r\n";
            $head .= $refererline;
            $head .= $authline;
            $head .= "\r\n";

            fputs($fp, $head);      
            while(!feof($fp) or ($eoheader==true)) {
                if($header=fgets($fp, 1024)) {
                    if ($header == "\r\n") {
                        $eoheader = true;
                        break;
                    } else {
                        $header = trim($header);
                    }

                    if($format == 1) {
                    $key = array_shift(explode(':',$header));
                        if($key == $header) {
                            $headers[] = $header;
                        } else {
                            $headers[$key]=substr($header,strlen($key)+2);
                        }
                    unset($key);
                    } else {
                        $headers[] = $header;
                    }
                }
            }
            return $headers;

        } else {
            return false;
        }
    }
?>

Regards.
Donovan

(10-May-2006 12:10)

If you want to get headers that current PHP process is going to send back to browser, see headers_list()

stuart at sixletterwords dot com (14-Sep-2005 08:52)

hey, i came across this afew weeks ago and used the function in an app for recording info about domains that my company owns, and found that the status this returns was wrong most of the time (400 bad request or void for sites that were clearly online). then looking into it i noticed the problem was that it wasn't able to get the correct info about sites with redirections. but thats not the full problem because everything on my server was returning the wrong status too. i searched around on php.net for other info and found that fsockopen's example worked better and only needed some tweeking.

heres the function i put together from it and a small change.

<?php
if(!function_exists('get_headers')) {
   function get_headers($url,$format=0,$httpn=0){
    $fp = fsockopen($url, 80, $errno, $errstr, 30);
    if ($fp) {
       $out = "GET / HTTP/1.1\r\n";
       $out .= "Host: $url\r\n";
       $out .= "Connection: Close\r\n\r\n";
       fwrite($fp, $out);
       while (!feof($fp)) {
           $var.=fgets($fp, 1280);
       }

    $var=explode("<",$var);
    $var=$var[0];
    $var=explode("\n",$var);
    fclose($fp);
    return $var;
    }
    }
}
?>

this returns an array of the header (only problem being that if the site doesn't have correct html it'll pull in some content too).

hope this'll help someone else.

sey at sey dot prometheus-designs dot net (28-Jul-2005 12:10)

The replacement updated get_headers function by aeontech at gmail dot com improperly formats dates when $format = 1.

Replace:
<?
else {
$headers[strtolower($h2[0])] = trim($h2[1]);
}
?>

With:
<?
else {   
$foo = implode( ':', $h2 );
$foo = preg_replace( '/[a-zA-Z- ]*: /', '', $foo );
$headers[strtolower($h2[0])] = trim( $foo );
}

drfickle2 at yahoo dot com (27-Jul-2005 01:01)

aeontech, this the below change adds support for SSL connections. Thanks for the code!

        if (isset($url_info['scheme']) && $url_info['scheme'] == 'https') {
            $port = 443;
            $fp=fsockopen('ssl://'.$url_info['host'], $port, $errno, $errstr, 30);
        } else {
           $port = isset($url_info['port']) ? $url_info['port'] : 80;
            $fp=fsockopen($url_info['host'], $port, $errno, $errstr, 30);
        }

(15-Jul-2005 09:37)

For anyone reading the previous comments, here is code that takes into account all the previous suggestions and includes a bugfix, too.

This code basically provides the "get_headers" function even on systems that are not running PHP 5.0.  It uses strtolower() on the keys, as suggested.  It uses the $h2 array instead of the $key, as suggested.  It removes a line about unsetting the $key -- no reason to unset something which is no longer used.  And I've changed the status header to be named "status" (instead of "0") in the array.  Note that if more than one header is returned without a label, they'll be stuck in "status" -- but I think status is the only header that comes back without a label, so it works for me.  So, first the code, then a sample of the usage:

<?php
if(!function_exists('get_headers')) {
   /**
   * @return array
    * @param string $url
    * @param int $format
    * @desc Fetches all the headers
    * @author cpurruc fh-landshut de
    * @modified by dotpointer
    * @modified by aeontech
    */
    function get_headers($url,$format=0) {
        $url_info=parse_url($url);
        $port = isset($url_info['port']) ? $url_info['port'] : 80;
        $fp=fsockopen($url_info['host'], $port, $errno, $errstr, 30);
        if($fp) {
            $head = "HEAD ".@$url_info['path']."?".@$url_info['query'];
            $head .= " HTTP/1.0\r\nHost: ".@$url_info['host']."\r\n\r\n";
            fputs($fp, $head);
            while(!feof($fp)) {
                if($header=trim(fgets($fp, 1024))) {
                    if($format == 1) {
                        $h2 = explode(':',$header);
                        // the first element is the http header type, such as HTTP/1.1 200 OK,
                        // it doesn't have a separate name, so we have to check for it.
                        if($h2[0] == $header) {
                            $headers['status'] = $header;
                        }
                        else {
                            $headers[strtolower($h2[0])] = trim($h2[1]);
                        }
                    }
                    else {
                        $headers[] = $header;
                    }
                }
            }
            return $headers;
        }
        else {
            return false;
        }
    }
}
?>

OK?  Here's the usage:

<?php
$response = get_headers('http://www.example.com/', 1);
if (!$response) {
    echo 'Unable to initiate connection.';
}
else {
    print_r($response);
}
?>

Chortos-2 (03-Jun-2005 10:44)

aeontech, I'd edit your function a little... How about replacing this:

<?
                       $key = array_shift(explode(':',$header));
                       // the first element is the http header type, such as HTTP 200 OK,
                       // it doesn't have a separate name, so we have to check for it.
                       if($key == $header)
                       {
                           $headers[] = $header;
                       }
                       else
                       {
                           $headers[$key]=substr($header,strlen($key)+2);
                       }
?>

with this:

<?
                       $h2 = explode(':',$header);
                       // the first element is the http header type, such as HTTP/1.1 200 OK,
                       // it doesn't have a separate name, so we have to check for it.
                       if($h2[0] == $header)
                       {
                           $headers[] = $header;
                       }
                       else
                       {
                           $headers[ $h2[0] ] = trim($h2[1]);
                       }
?>

I think it looks a bit nicer :)

david at nothisbit dot futuresbright dot com (05-Apr-2005 12:03)

After discovering that some webservers reply with "Content-Type" and others with "Content-type" I modified the function below to use strtolower($key) to make for easy checking against these case differences.

aeontech at gmail dot com (24-Dec-2004 01:20)

In response to dotpointer's modification of Jamaz' solution...

Here is a small modification of your function, this adds the emulation of the optional $format parameter.

<?php
if(!function_exists('get_headers')) {
   
    /**
    * @return array
    * @param string $url
    * @param int $format
    * @desc Fetches all the headers
    * @author cpurruc fh-landshut de
    * @modified by dotpointer
    * @modified by aeontech
    */
    function get_headers($url,$format=0)
    {
        $url_info=parse_url($url);
        $port = isset($url_info['port']) ? $url_info['port'] : 80;
        $fp=fsockopen($url_info['host'], $port, $errno, $errstr, 30);
       
        if($fp)
        {
            $head = "HEAD ".@$url_info['path']."?".@$url_info['query']." HTTP/1.0\r\nHost: ".@$url_info['host']."\r\n\r\n";      
            fputs($fp, $head);      
            while(!feof($fp))
            {
                if($header=trim(fgets($fp, 1024)))
                {
                    if($format == 1)
                    {
                        $key = array_shift(explode(':',$header));
                        // the first element is the http header type, such as HTTP 200 OK,
                        // it doesn't have a separate name, so we have to check for it.
                        if($key == $header)
                        {
                            $headers[] = $header;
                        }
                        else
                        {
                            $headers[$key]=substr($header,strlen($key)+2);
                        }
                        unset($key);
                    }
                    else
                    {
                        $headers[] = $header;
                    }
                }
            }
            return $headers;
        }
        else
        {
            return false;
        }
    }
}
?>