Redirecting and Remapping with mod_rewrite

Assume we have recently renamed the page foo.html to bar.html and now want to provide the old URL for backward compatibility. However, we want that users of the old URL even not recognize that the pages was renamed - that is, we don't want the address to change in their browser.

We rewrite the old URL to the new one internally via the following rule:

RewriteEngine  on
RewriteRule    ^/foo\.html$  /bar.html [PT]

Assume again that we have recently renamed the page foo.html to bar.html and now want to provide the old URL for backward compatibility. But this time we want that the users of the old URL get hinted to the new one, i.e. their browsers Location field should change, too.

We force a HTTP redirect to the new URL which leads to a change of the browsers and thus the users view:

RewriteEngine  on
RewriteRule    ^/foo\.html$  bar.html  [R]

In this example, as contrasted to the internal example above, we can simply use the Redirect directive. mod_rewrite was used in that earlier example in order to hide the redirect from the client:

Redirect /foo.html /bar.html

If a resource has moved to another server, you may wish to have URLs continue to work for a time on the old server while people update their bookmarks.

You can use mod_rewrite to redirect these URLs to the new server, but you might also consider using the Redirect or RedirectMatch directive.

#With mod_rewrite
RewriteEngine on
RewriteRule   ^/docs/(.+)  http://new.example.com/docs/$1  [R,L]

#With RedirectMatch
RedirectMatch ^/docs/(.*) http://new.example.com/docs/$1

#With Redirect
Redirect /docs/ http://new.example.com/docs/

How can we transform a static page foo.html into a dynamic variant foo.cgi in a seamless way, i.e. without notice by the browser/user.

We just rewrite the URL to the CGI-script and force the handler to be cgi-script so that it is executed as a CGI program. This way a request to /~quux/foo.html internally leads to the invocation of /~quux/foo.cgi.

RewriteEngine  on
RewriteBase    /~quux/
RewriteRule    ^foo\.html$  foo.cgi   [H=cgi-script]

How can we make URLs backward compatible (still existing virtually) after migrating document.YYYY to document.XXXX, e.g. after translating a bunch of .html files to .php?

We rewrite the name to its basename and test for existence of the new extension. If it exists, we take that name, else we rewrite the URL to its original state.

#   backward compatibility ruleset for
#   rewriting document.html to document.php
#   when and only when document.php exists
<Directory /var/www/htdocs>
    RewriteEngine on
    RewriteBase /var/www/htdocs

    RewriteCond $1.php -f
    RewriteCond $1.html !-f
    RewriteRule ^(.*).html$ $1.php
</Directory>

This example uses an often-overlooked feature of mod_rewrite, by taking advantage of the order of execution of the ruleset. In particular, mod_rewrite evaluates the left-hand-side of the RewriteRule before it evaluates the RewriteCond directives. Consequently, $1 is already defined by the time the RewriteCond directives are evaluated. This allows us to test for the existence of the original (document.html) and target (document.php) files using the same base filename.

This ruleset is designed to use in a per-directory context (In a <Directory> block or in a .htaccess file), so that the -f checks are looking at the correct directory path. You may need to set a RewriteBase directive to specify the directory base that you're working in.

The very best way to solve this doesn't involve mod_rewrite at all, but rather uses the Redirect directive placed in a virtual host for the non-canonical hostname(s).

<VirtualHost *:80>
  ServerName undesired.example.com
  ServerAlias example.com notthis.example.com

  Redirect / http://www.example.com/
</VirtualHost>

<VirtualHost *:80>
  ServerName www.example.com
</VirtualHost>

You can alternatively accomplish this using the <If> directive:

<If "%{HTTP_HOST} != 'www.example.com'">
	Redirect / http://www.example.com/
</If>

Or, for example, to redirect a portion of your site to HTTPS, you might do the following:

<If "%{SERVER_PROTOCOL} != 'HTTPS'">
	Redirect /admin/ https://www.example.com/admin/
</If>

If, for whatever reason, you still want to use mod_rewrite - if, for example, you need this to work with a larger set of RewriteRules - you might use one of the recipes below.

For sites running on a port other than 80:

RewriteCond %{HTTP_HOST}   !^www\.example\.com [NC]
RewriteCond %{HTTP_HOST}   !^$
RewriteCond %{SERVER_PORT} !^80$
RewriteRule ^/?(.*)         http://www.example.com:%{SERVER_PORT}/$1 [L,R,NE]

And for a site running on port 80

RewriteCond %{HTTP_HOST}   !^www\.example\.com [NC]
RewriteCond %{HTTP_HOST}   !^$
RewriteRule ^/?(.*)         http://www.example.com/$1 [L,R,NE]

If you wanted to do this generically for all domain names - that is, if you want to redirect example.com to www.example.com for all possible values of example.com, you could use the following recipe:

RewriteCond %{HTTP_HOST} !^www\. [NC]
RewriteCond %{HTTP_HOST} !^$
RewriteRule ^/?(.*) http://www.%{HTTP_HOST}/$1 [L,R,NE]

These rulesets will work either in your main server configuration file, or in a .htaccess file placed in the DocumentRoot of the server.

A particular resource might exist in one of several places, and we want to look in those places for the resource when it is requested. Perhaps we've recently rearranged our directory structure, dividing content into several locations.

The following ruleset searches in two directories to find the resource, and, if not finding it in either place, will attempt to just serve it out of the location requested.

RewriteEngine on

#   first try to find it in dir1/...
#   ...and if found stop and be happy:
RewriteCond         %{DOCUMENT_ROOT}/dir1/%{REQUEST_URI}  -f
RewriteRule  ^(.+)  %{DOCUMENT_ROOT}/dir1/$1  [L]

#   second try to find it in dir2/...
#   ...and if found stop and be happy:
RewriteCond         %{DOCUMENT_ROOT}/dir2/%{REQUEST_URI}  -f
RewriteRule  ^(.+)  %{DOCUMENT_ROOT}/dir2/$1  [L]

#   else go on for other Alias or ScriptAlias directives,
#   etc.
RewriteRule   ^  -  [PT]

We have numerous mirrors of our website, and want to redirect people to the one that is located in the country where they are located.

Looking at the hostname of the requesting client, we determine which country they are coming from. If we can't do a lookup on their IP address, we fall back to a default server.

We'll use a RewriteMap directive to build a list of servers that we wish to use.

HostnameLookups on
RewriteEngine on
RewriteMap    multiplex         txt:/path/to/map.mirrors
RewriteCond  %{REMOTE_HOST}     ([a-z]+)$ [NC]
RewriteRule   ^/(.*)$  ${multiplex:%1|http://www.example.com/}$1  [R,L]

The RewriteCond directive captures the last portion of the hostname of the requesting client - the country code - and the following RewriteRule uses that value to look up the appropriate mirror host in the map file.

We wish to provide different content based on the browser, or user-agent, which is requesting the content.

We have to decide, based on the HTTP header "User-Agent", which content to serve. The following config does the following: If the HTTP header "User-Agent" contains "Mozilla/3", the page foo.html is rewritten to foo.NS.html and the rewriting stops. If the browser is "Lynx" or "Mozilla" of version 1 or 2, the URL becomes foo.20.html. All other browsers receive page foo.32.html. This is done with the following ruleset:

RewriteCond %{HTTP_USER_AGENT}  ^Mozilla/3.*
RewriteRule ^foo\.html$         foo.NS.html          [L]

RewriteCond %{HTTP_USER_AGENT}  ^Lynx/ [OR]
RewriteCond %{HTTP_USER_AGENT}  ^Mozilla/[12]
RewriteRule ^foo\.html$         foo.20.html          [L]

RewriteRule ^foo\.html$         foo.32.html          [L]

On some webservers there is more than one URL for a resource. Usually there are canonical URLs (which are be actually used and distributed) and those which are just shortcuts, internal ones, and so on. Independent of which URL the user supplied with the request, they should finally see the canonical one in their browser address bar.

We do an external HTTP redirect for all non-canonical URLs to fix them in the location view of the Browser and for all subsequent requests. In the example ruleset below we replace /puppies and /canines by the canonical /dogs.

RewriteRule   ^/(puppies|canines)/(.*)    /dogs/$2  [R]

RedirectMatch ^/(puppies|canines)/(.*) /dogs/$2

Usually the DocumentRoot of the webserver directly relates to the URL "/". But often this data is not really of top-level priority. For example, you may wish for visitors, on first entering a site, to go to a particular subdirectory /about/. This may be accomplished using the following ruleset:

We redirect the URL / to /about/:

RewriteEngine on
RewriteRule   ^/$  /about/  [R]

Note that this can also be handled using the RedirectMatch directive:

RedirectMatch ^/$ http://example.com/about/

Note also that the example rewrites only the root URL. That is, it rewrites a request for http://example.com/, but not a request for http://example.com/page.html. If you have in fact changed your document root - that is, if all of your content is in fact in that subdirectory, it is greatly preferable to simply change your DocumentRoot directive, or move all of the content up one directory, rather than rewriting URLs.

As of version 2.2.16, you should use the FallbackResource directive for this:

<Directory /var/www/my_blog>
  FallbackResource index.php
</Directory>

However, in earlier versions of Apache, or if your needs are more complicated than this, you can use a variation of the following rewrite set to accomplish the same thing:

<Directory /var/www/my_blog>
  RewriteBase /my_blog

  RewriteCond /var/www/my_blog/%{REQUEST_FILENAME} !-f
  RewriteCond /var/www/my_blog/%{REQUEST_FILENAME} !-d
  RewriteRule ^ index.php [PT]
</Directory>

If, on the other hand, you wish to pass the requested URI as a query string argument to index.php, you can replace that RewriteRule with:

RewriteRule (.*) index.php?$1 [PT,QSA]

Note that these rulesets can be used in a .htaccess file, as well as in a <Directory> block.

Many of the solutions in this section will all use the same condition, which leaves the matched value in the %2 backreference. %1 is the beginining of the query string (up to the key of intererest), and %3 is the remainder. This condition is a bit complex for flexibility and to avoid double '&&' in the substitutions.

• This solution removes the matching key and value:

  # Remove mykey=???
  RewriteCond %{QUERY_STRING} (.*(?:^|&))mykey=([^&]*)&?(.*)&?$
  RewriteRule (.*) $1?%1%3

• This solution uses the captured value in the URL subsitution, discarding the rest of the original query by appending a '?':

  # Copy from query string to PATH_INFO
  RewriteCond %{QUERY_STRING} (.*(?:^|&))mykey=([^&]*)&?(.*)&?$
  RewriteRule (.*) $1/products/%2/? [PT]

• This solution checks the captured value in a subsequent condition:

  # Capture the value of mykey in the query string
  RewriteCond %{QUERY_STRING} (.*(?:^|&))mykey=([^&]*)&?(.*)&?$
  RewriteCond %2 !=not-so-secret-value 
  RewriteRule (.*) - [F]

• This solution shows the reverse of the previous ones, copying path components (perhaps PATH_INFO) from the URL into the query string.

  # The desired URL might be /products/kitchen-sink, and the script expects 
  # /path?products=kitchen-sink.
  RewriteRule ^/?path/([^/]+)/([^/]+) /path?$1=$2 [PT]