Hi Richard

Thanks for your reply. Google-CSE makes definetly sense, as long as the data is 
publicly accessible.

In my case the content is only accessible after a user logs in. Depending on 
the user role there are different data and dms nodes, that the search shouldn't 
find. As far as I know the magnolia search takes this into consideration.

How did you combine the two scores? We decided to create 3 tabs, one per 
repository. Therefore I only have to combine 2 searches for each repo (due to 2 
different searchpaths).

Thanks for all your inputs!
Armin

 

Am 01.02.2013 um 14:35 schrieb "Unger, Richard" <[email protected]>:

> Hello Will, hello Armin,
> 
> A lot of time has passed since that original post of mine.
> 
> Since then we have gone away from using the JCR search. Instead we use 
> google-CSE, for a number of reasons:
> 
> - google is better at indexing. Google searches all kinds of content, 
> including PDFs and other common file types, and finds good matches for 
> everything
> - google combines results - documents and web-content get indexed and scored 
> in the same way
> - google only searches public data - since the CSE scans from the outside, 
> there is no chance of "secrets" landing in the index
> - google permits more powerful queries - by using the different query filters 
> and providing sitemaps with extra data we can define all kinds of very 
> specific searches (eg: search for the term "food" within the path 
> "/some/path" returning documents in English with the data property 
> "important")
> 
> --> while many of these examples might also be achievable with JCR Searches 
> you would have to do quite some tweaking of the indexing configuration.
> 
> In the JCR Search, we originally simply combined the 2 result-lists into one, 
> using the scores as they came. The result wasn't bad, seemed to make sense.
> Reading into it in some more detail after you pointed out the problem with 
> this approach showed that whether this works out or not is more a matter of 
> luck than anything else.
> In JCR 1 it simply isn't possible to combine search results because of the 
> scoring. In JCR 2.0, I have not checked if there is any difference with 
> respect to this problem...
> 
> Regards from Vienna,
> 
> Richard
> 
> 
> 
> -----Ursprüngliche Nachricht-----
> Von: [email protected] 
> [mailto:[email protected]] Im Auftrag von Will Scheidegger 
> (via Magnolia Forums)
> Gesendet: Mittwoch, 30. Jänner 2013 11:40
> An: Magnolia User List
> Betreff: [magnolia-user] Re: Magnolia search on multiple repositories
> 
> Hm... yes, still interested in a good answer myself!
> 
> Richard, would you mind describing your step "3. Combine the results, sort by 
> score." in more details?
> 
> Thanks!
> -will
> 
> -- 
> Context is everything: 
> http://forum.magnolia-cms.com/forum/thread.html?threadId=a351c58e-1ef7-4a6a-94ad-fd6ea99ee583
> 
> 
> ----------------------------------------------------------------
> For list details, see http://www.magnolia-cms.com/community/mailing-lists.html
> Alternatively, use our forums: http://forum.magnolia-cms.com/
> To unsubscribe, E-mail to: <[email protected]>
> ----------------------------------------------------------------
> 
> 
> 
> 
> 
> ----------------------------------------------------------------
> For list details, see http://www.magnolia-cms.com/community/mailing-lists.html
> Alternatively, use our forums: http://forum.magnolia-cms.com/
> To unsubscribe, E-mail to: <[email protected]>
> ----------------------------------------------------------------
> 



----------------------------------------------------------------
For list details, see http://www.magnolia-cms.com/community/mailing-lists.html
Alternatively, use our forums: http://forum.magnolia-cms.com/
To unsubscribe, E-mail to: <[email protected]>
----------------------------------------------------------------

Reply via email to