How do you prevent duplicate URLs for two pages that have the same title, but are located in different folders in Author?

Hi @ozziesa 

Below is my understanding from the above question:

You have a page as:

/content/project/07/2021/someurl.html

/content/project/08/2021/someurl.html

So both the URLs can reside in author and publish. But when you will shorten the month and year, the URL will look something like to end user:

/content/project/someurl.html

/content/project/someurl.html

If you see above both the URLs are same and when user will request the URL, dispatcher will not be able to identify to request resource from which URL. So the URL needs to be unique that means it cannot reside in same domain when the shortening is enabled or you might need to handle it based on query parameter where the URL will become something like below and using the redirects at the dispatcher it can be managed.

/content/project/someurl.html?m=07

/content/project/someurl.html?m=08

RewriteCond %{REQUEST_URI} ^/content/project/someurl.html
RewriteCond %{QUERY_STRING} m="07"
RewriteRule (.*) /content/project/07/2021/someurl.html [R=302,L]

RewriteCond %{REQUEST_URI} ^/content/project/someurl.html
RewriteCond %{QUERY_STRING} m="08"
RewriteRule (.*) /content/project/08/2021/someurl.html [R=302,L]

If you do not want to keep both the pages live on publish and want to prevent publication of the 2nd URL when the first URL is live, then you need to write a custom ReplicationPreprocessor which will be invoked each time you trigger a replication request. The logic can be to check the current page name(i.e., someurl here) and look for any occurances of the same page within the content hirearchy and if found check for the replication status of that page, if the page is replicated then stop the replication of current page, if not then proceed with replication. This will ensure only one copy is present in publish and you do not have to manage any redirects.

Here is a sample of Preprocessor:

https://labs.tadigital.com/index.php/2019/06/25/aem-preprocessor/

Hope this helps!

Thanks 

Hi,

you cannot map an incoming request to multiple AEM path. But what you can do is serve a page from different folder by intercepting the request using sling filter.

Sling Filter will check the request and serve the correct or latest version of page.but you have to build a mechanism to clear the cache, when there is new version of the page is availble.

if you are using ACS common then it is easy to delete cache by pubishing other content.

Hello @ozziesa ,

A single url path will not be able to resolve to two AEM content paths. See https://www.aemquickstart.in/2015/11/how-is-resource-resolution-done-in-sling_85.html  for more info. We can either try to use query parameters or selectors (preferred since this will be cached) to take the users to different page locations.

I can think of following approach:

Considering ,how are the end users going to see these pages. If the links are supposed to be showing only in a particular component or a few, we can use a common service/Util method to generate the link to go to the desired page.

We can have the links of pages generated based on the parent structure and use of selectors

For eg: say we have "/content/project/en/2021/08/titlename.html" and "/content/project/en/2021/07/titlename.html"

The backend of the component will create the display link of the above links as "/content/project/en.uniquesel.2021.08.titlename.html" and "/content/project/en.uniquesel.2021.07.titlename.html"

Then we can use a Sling Servlet Filter (Example) that will read the selectors, do a quick check whether the request url at [0] contains/equals 'uniquesel' and proceed. Further, create the destination path in the filter based on the request url selectors other than [0]

i.e "/content/project/en.uniquesel.2021.08.titlename.html" --> "/content/project/en/2021/08/titlename.html"

Hence redirecting the user to the desired destination.

The page structure for this may vary in reality say /content/project/en/2021/08/01/titlename.html and /content/project/en/2021/07/31/titlename.html. We will have to think of or consider what kind of page link the end user may possibly be.

Not sure if this serves the purpose, but the answer to question 'how important is the requirement is?' vs 'loe and adding complexity' must be considered.

Regarding your initial question: This does not work properly, and the reasons for it has been described already by the other responses.

Also, I find it odd that you create a content structure which you are then trying to hide. The content/information structure is an inherently important piece in your AEM application design, and has impact to many aspects, see [1]. 

To your second question: You should not do that and restrict authors from naming pages as they need.

[1] https://cqdump.joerghoh.de/2017/11/13/creating-the-content-architecture-with-aem/