Oracle Text v/s Lucene
528121Jun 5 2007 — edited Jun 23 2007Hi -
We are currently looking into adding full text search capabilities to our site.
I am trying to compile a list that discusses the relative pros and cons about Oracle Text versus file-system based search engines such as Lucene.
Here's our configuration
* RDBMS is Enterprise Oracle 10g
* Dual fiber channel RAID-5 configuration
* 2-node cluster
* 8GB RAM/per node
* Dual 3.6GHz Intel CPU/per node
* 100% of the content to be indexed is stored in our RDBMS
* Current table size 3 Million records, can go up to 100 million records
Since we have the Oracle Enterprise license, cost is not an issue (oracle text comes with it).
Some anecdotal evidence I have heard thus far:
* speed - lucene is faster
* throughput – adding more load requires additional DB CPU licenses, which is far more expensive than cheaper app server cpus and memory
* RDBMS server off load (someone reported they offloaded 70% of db server)
* index size - lucene indexes are smaller
* Scoring algorithm: Not possible to customize with Oracle Text
* Progressive Relaxation: Adding expansion terms such as Fuzzy, Synonym, Near slows search down considerably.
Does anyone have any real-world production experience with Lucene and/or Oracle Text? Thanks in advance.