My URL structure looks a bit like this:

  • /group1/
  • /group2/
  • /group3/

Group3 (and it's subpages) are also available in another language at the following URLs. A lang prefix was added to this URL:

  • /lang/group3/

Unfortunately, a CMS upgrade has caused the following URLs to now considered valid, even though they have not been translated.

  • /lang/group1
  • /lang/group2

I would like to allow /group3 and /lang/group3 but block /lang/group* from the index.

What is the correct robots syntax to do this?

1 Answer 1


If you literally only have a few "groups" you want to block then you would do something like:

User-agent: *
Disallow: /lang/group1
Disallow: /lang/group2

...and everything else would be allowed. This would work with all robots that obey the original "standard". Or, you could block all groups (group1, group2, etc.) and make an exception for "group3", like:

User-agent: *
Disallow: /lang/group
Allow: /lang/group3

Note that the Allow directive is not part of the original "standard", but has universal support. The URL path is simply a prefix.

HOWEVER, I wouldn't use robots.txt to block the pages being "crawled". What about stray visitors? And bad bots? And robots.txt doesn't prevent pages from being indexed if they are inadvertently linked to. I would use .htaccess or your server config to actually block all traffic to these URLs. Something like the following in .htaccess:

RewriteEngine On
RewriteRule ^lang/group[12] - [R=404]

To respond with a 404 for all requests to these invalid URLs. Or, use the F flag to respond with a 403 "Forbidden".

  • Thanks, there are many thousands 'groups' I'm afraid. I agree .htaccess would be better. Would a pattern like Disallow: /lang/* Allow: /lang/group3 be considered valid?
    – Tom
    Jan 12, 2015 at 12:22
  • 1
    Yes that would be valid, although the * at the end of the URL path is unnecessary (and a modern addition to the standard). The URL path is just a prefix anyway.
    – MrWhite
    Jan 12, 2015 at 12:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.