Question : index is not being used if not forced..why?

Hi,
I have one table named "resources"
which has 2 nonclustered indexes
1. 1 composite on "dep_num" and "date" , name - idx_date_name
2. 1 on single on "id" - idx_id

when i use the query as

select * from resources where id='xyz'  sybase optimizer is ignoring the index 2.
where as if I use the force index option as in
select * from resources (index idx_id) where id='xyz'

it uses the  index properly. what could be the problem? the table has around 15 mn rows.

Answer : index is not being used if not forced..why?

Clustered indexes can be non-unique, but there is a performance penalty in doing so. 9000 rows out of 15 million might not be too bad. It'll be an expensive (time, disk space) process to create one for testing purposes, though.

Is there a second column frequently used in queries that could be added to id to make it a composite unique index?

When I hear a similar table is behaving in another database, I immediately suspect fragmentation issues, especially given this is a heap table (no clustered index). Is it a DOL table (datapages or datarows)? If so, consider running a "reorg rebuild" - also slow and there are some prerequisites.

I'm still curious about index statistics. This will take a while, but try:

delete statistics resources
go
update index statistics resources using 100 values
go

This will generate better statistics than the default. Worth trying.

And...

>> select * from resources where id='xyz'
>
>Table: Inventory  scan count 1,  logical

If that isn't a typo or a cut & paste issue, the table names don't match?