Search with Hibernate ignites some questions in my mind:
1- How can we distinguish search using Apache Lucene and with Hibernate? 2- Is it fast if we compare this with apache lucene or SOLR on the same matrices? 3- Can we integrate hibernate with apache lucene to make it work better?
Congrats Emmanuel Bernard & John Griffin coming up with a good book, though don't know much about.
3. This is exactly what Hibernate Search does. It uses Lucene as its engine but abstracts all the low level work you must do when using Lucene alone. In addition we provide clustered searching out of the box with some simple configuration options. We also solve the 'pessimistic lock' update problem of Lucene. There are many more benefits too numerous to mention.
2. I cannot speak for SOLR. I haven't used it but we are as fast as Lucene is when querying and ease of use of Hibernate Search should take it 'over the top' with implementors.
1. I'm not quite sure I know what you are asking with this question. Can you be a little more specific?
Joined: Feb 28, 2007
I got answer to my first question, as you said It uses Lucene as its engine but abstracts all the low level work you must do when using Lucene alone.
Relevancy scoring configuration should be more easy in configuration, and customized as I hope it does with Hibernate.
How easy and customizable will it be configuring relevancy scoring? [ December 08, 2008: Message edited by: Chandra Bhatt ]
Joined: Oct 24, 2008
Changing relevance scoring is fairly easy. Getting it correct is another story. I wrote chapter 12 which deals exclusively with scoring modification and can be an intense read. It covers overriding the DefaultSimilarity class and writing your own Weight and Score classes.
The key thing to remember is changing it is easy, getting at you want is hard. You will have to test every little change to make sure it's what you want.