Originally posted by rsuri
Thanks for your response guys.
We have a developer who when learned about bitmap indexes, started converting every index into a bitmap - his only logic being that they were much smaller than the b*tree and the performance was about the same if not better than b*tree.
Evaluating the overhead/trouble we are getting into we sure do need to do some benchmarks.

Thanks,

-Rajeev
I think that what your developer needs is some discipline imposed on him/her/it.

The tendency to grab hold of any shiny new features that the database offers without fully investigating the pro's and con's (like reading the doc's, searching metalink for bugs, testing fully) gives this person the Mark Of the Idiot stamped hard on their forehead.

I wouldn't let them near a database, personally.