If i disallow unfriendly URL via robots.txt, will its friendly counterpart still be indexed?

[mrwestern]

Our not-so-lovely CMS loves to render pages regardless of the URL structure, just as long as the page name itself is correct. For example, it will render the following as the same page:

example.com/123.html

example.com/dumb/123.html

example.com/really/dumb/duplicative/URL/123.html

To help combat this, we are creating mod rewrites with friendly urls, so all of the above would simply render as example.com/123

I understand robots.txt respects the wildcard (*), so I was considering adding this to our robots.txt:

Disallow: */123.html

If I move forward, will this block all of the potential permutations of the directories preceding 123.html yet not block our friendly example.com/123?

Oh, and yes, we do use the canonical tag religiously - we're just mucking with the robots.txt as an added safety net.

[irvingw]

that disallow command will block all files with the name 123.html in any folder deeper that the root.

This with the canonical (absolute not relative) will probably cover you, but it is really recommended to get a robots noindex meta tag on these duplicate pages as well. Bots coming in from an external link pointing to that page could result in the page getting indexed, also the canonical is a suggestion not a rule.

[mrwestern]

Thanks, however, the meta tag won't work in this case because it's technically one page with an infinite amount of names via the URL (remember, the CMS only depends on the 123.html and ignores the directories preceding it). If I applied the NOINDEX within the meta, then the version I do want to get indexed would not be indexed.

The question was really around "will the internal rewrite of /123.html to just /123 be impacted if we disallow */123.html" - and since the rewrite happens before the bot sees it, I presume the answer is "no, it will not be impacted: 123.html will be blocked yet /123 will still be indexed.

Now, after I posted the question I realized this is the case where I should use a "greedy" 301 redirect via htaccess rather than try to block permutations of the URL via robots.txt. So I decided to not go the robots.txt route and instead do a 301 redirect via regex:

*/123.html to /123 (that's obviously not perfect regex, but you see my point)

[Cyrus-Shepard]

Yeah, if you could solve this via .htaccess that would be great, especially if you have link equity flowing into any of those URLs.

I'd go one step further than Irving and highly recommend canonical tags on those URLs. Since, as you said, it's all one page with infinite URL possibilities, the canonical should be easy to implement.

Best of luck!