AnsweredAssumed Answered

Alfresco 3.2:Problem with Lucene indexing

Question asked by sergiofigueras on Aug 20, 2010
Hi everyone,

I'm using Alfresco for a simple and little repository, around 70k documents. I made a webservice class wich sends the documents to the Alfresco. Ok, it's working fine.

So, I need to put a set of keys to the user search and get tthe results. I've modified the Content model and added a property named by cm:Data , wich is a string that contains things like "type of the document key1 key2". I added the index enabled true, tokenised both, atomic true to this property. Remembering: I've changed the default contentModel.

Ok, I upload the documents in different spaces.

But… when I use the Store Node and lucene search, looking by example for: "TYPE:"cm:content" @cm\:Data:TYPE OF THE DOC* ", it results a different number of the correct number that it should return. For example, there are seven docs in repository, it returns a single node, for example.

Looking for answers, I found a property in repository.properties I got:
lucene.maxAtomicTransformationTime=20

I've changed that to 2000, but, still I got no correct answer.

And if I rebuild all the lucene indexes, the result changes. Givin' me around 5% of the correct results…

Please consider somethings:
-There's no problem with my upload code. It's really uploading the properties correctly. And setting the cm:Data property to d:text type. So, Lucene should index it.
- There's no problem with Lucene query formation.
- I'm using Ubuntu 8.04 server.

Important note:
- I'm not uploading a byte array to Alfresco. I'm just uploading some properties and saving a null byte array. It doesn't give me any error at all.

Thanks!

Sergio.

Outcomes