7

I am accessing some Dynamic Component Presentations from Tridion in my website. I have more than one website (built in Java) on the same Tridion server.

Now I want o include search functionality in the website that searches these Dynamic Component Presentations and fetches the results.

I have got some info on SOLR search engine.

I do not know if I have got the right one and how complex it would be.

Could someone help me to implement search functionality for these Dynamic Component Presentations?

Improve this question

4 Answers 4

Reset to default
5

Instead of indexing the DB, maybe the best way would be to index directly the final web site itself to generate the search index.

As SOLR does not include a crawler, you will have to combine it with an external crawler like Nutch.

Improve this answer
6

If all you want is search by metadata - use the SDL Tridion Criteria APIs.

If you want to do full text search, then you probably want more than just search on DCPs, and you should consider doing one of the approaches suggested on this post.

Meanwhile, there is also this great effort done by Will Price and Raimond Kempees: SI4T.

Improve this answer
6

If you do decide to use SOLR for site search (which I recommend,) consider a custom deployer / storage extension approach (which I believe SI4T does). It is much more efficient that crawling with Nutch especially if the site is large.

  • Items are indexed at publish time, and therefore are immediately visible in search.
  • Items are removed from the index at un-publish time and therefore immediately removed from search

It is a little more complex than indexing with Nutch, but will pay off in the future

Improve this answer
1
  • The question is specifically about dynamic component presentations, which might make a crawling solution better for some designs. e.g. how easy is it to map the content to a URL where it can be reached?
    – Dominic Cronin
    Commented Oct 20, 2013 at 7:42
4

We have implemented SOLR for site search in a big implementation and believe me it is much more effective and fast as compared to Nutch - Still working in production without any issue.

From implementation perspective, it is very simple to implement as you may need to define few configuration in the XML and then query solr and handle the data return.

The architecture of this implementation is described HERE

Improve this answer

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.