Skip to Main Content

Database Software

Announcement

For appeals, questions and feedback about Oracle Forums, please email oracle-forums-moderators_us@oracle.com. Technical questions should be asked in the appropriate category. Thank you!

Context Index Search Is Slow

User_KOMEBJul 20 2009 — edited Jul 23 2009
I have a table that contains about 10 million rows, which contains a clob that stores about a page worth of text. I created a context index on the clob. But when I use the context index to search and rank the documents, it is really slow. Most times search take 1-5 minutes. Does this sound normal? I do notice that after I do an initial search, the next search is much faster. Not sure if that is becuase it cached everthing or not. If it is a caching issue, I'm not sure if it is the $I table or the table containing the clob text.

I have read the documentation and I am still a little unsure on how the internals of the index work. I know when I do a search, it finds the docid's in the $I table. Does it then take these docid's and pull the clob from the base document table? The $I table also has about 10 million rows. I have done a full optimization but I noticed that there are still multiple rows for many tokens in the $I table. It looks like the index doesn't like the blob in the $I table to be greater than 4K. I'm guessing this is done to keep the blob stored with the table. When I read the documentation, it says a full optimiation combines all docid's for a token into a single row. I'm not sure if this is part of my problem.

I have set up the index with all parameters at default. Is there any parameters that I could try to improve the performance of my index. The documents I am indexing are alot like these posts in these forums, but average about a page in length. Thanks for any help!
Comments
Locked Post
New comments cannot be posted to this locked post.
Post Details
Locked on Aug 20 2009
Added on Jul 20 2009
7 comments
3,730 views