URL rewrite with HTACCESS for multiple language support

Question

I am looking to add multiple language support to my website. Is it possible to use the .htaccess file to change something like:

• example.com/dir/?lang=en to example.com/en/dir/
• example.com/main/?lang=de to example.com/de/main/
• example.com/main/page.php?lang=de to example.com/de/main/page.php

Where this works with any possible directories - so if for instance later on I made a new directory, I wouldn't need to add to this. In the above example, I want the latter to be what the user types in/is in the address bar, and the start to be how it is used internally.

Answer 1

Is it possible to use the .htaccess file to change something like:

• example.com/dir/?lang=en to example.com/en/dir/

Yes, except that you don't change the URL to example.com/en/dir/ in .htaccess . You change the URL to example.com/en/dir/ in your internal links in your application, before you change anything in .htaccess . This is the canonical URL and is "what the user types in/is in the address bar" - as you say.

You then use .htaccess to internally rewrite the request from example.com/en/dir/ , back into the URL your application understands, ie. example.com/dir/?lang=en (or rather example.com/dir/index.php?lang=en - see below). This is entirely hidden from the user. The user only ever sees example.com/en/dir/ - even when they look at the HTML source.

So, we need to rewrite /<lang>/<url-path>/ to /<url-path>/?lang=<lang> . Where <lang> is assumed to be a 2 character lowercase language code. If you are offering only a small selection of languages then this should be explicitly stated to avoid conflicts. We can also handle any additional query string on the original request (if this is requried). eg. /<lang>/<url-path>/?<query-string> to /<url-path>/?lang=<lang>&<query-string> .

A slight complication here is that a URL of the form /dir/?lang=en is not strictly a valid endpoint and requires further rewriting. I expect you are relying on mod_dir to issue an internal subrequest for the DirectoryIndex, eg. index.php ? So, really, this should be rewritten directly to /dir/index.php?lang=en - or whatever the DirectoryIndex document is defined as.

For example, in your root .htaccess file:

RewriteEngine On

# Rewrite "/<lang>/<directory>/" to `/<directory>/index.php?lang=<lang>"
RewriteCond %{DOCUMENT_ROOT}/$2/index.php -f
RewriteRule ^([a-z]{2})/(.*?)/?$ $2/index.php?lang=$1 [L]

# Rewrite "/<lang>/<directory>/<file>" to `/<directory>/<file>?lang=<lang>"
RewriteCond %{DOCUMENT_ROOT}/$2 -f
RewriteRule ^([a-z]{2})/(.+) $2?lang=$1 [L]

If you have just two languages (as in your example), or a small subset of known languages then change the ([az]{2}) subpattern to use alternation and explicitly identify each language code, eg. (en|de|ab|cd) .

This does assume you don't have physical directories in the document root that consist of 2 lowercase letters (or match the specific language codes).

Only URLs where the destination directory (that contains index.php ) or file exists are rewritten.

This will also rewrite requests for the document root (not explicitly stated in your examples). eg. example.com/en/ (trailing slash required here) is rewritten to /index.php?lang=en .

The regex could be made slightly more efficient if requests for directories always contain a trailing slash. In the above I've assumed the trailing slash is optional, although this does potentially create a duplicate content issue unless you resolve this in some other way (eg. rel="canonical" link element). So, in the code above, both example.com/en/dir/ (trailing slash) and example.com/en/dir (no trailing slash) are both accessible and both return the same resource, ie. /dir/index.php?lang=en .