Re: TO_DATE function causes table scan

frank.van.bortel_at_gmail.com wrote:
> Don't be obsessed with the use of an index; how did you come to the
> conclusion your query will be faster when it uses the index?
> 3500 rows is nothing, and tells nothing; what could tell something is
> which portion of the 3500 rows meets your criteria; CBO used to switch
> to FTS when more than 5-30% of the total records would be fetched.
> Also, values of optimizer_index_cost_adjust, optimizer_index_caching
> and db_file_multiblock_read_count are of importance.
>
> A good document on that would be "The Search for Intelligent Life in
> the Cost-Based Optimizer" on
> http://www.evdbt.com/SearchIntelligenceCBO.doc

I hadn't noticed the 3500 row figure before. But given that this likely fits into a very small number of blocks it is quite probable that the index is a waste of a perfectly good 4K block.

-- 
Daniel A. Morgan
http://www.psoug.org
damorgan_at_x.washington.edu
(replace x with u to respond)