Re: Index management

As an aside, Burleson has an article on whether to rebuild indexes at www.dbazine.com. It's an interesting read, largely because it comes to no conclusions (which is progress from him).

The biggest chuckle I got was the line "For now, the most any Oracle professional can do is to explore their indexes and learn how the software manages to [sic] b-tree structures."

I couldn't agree more, and I can think of at least one "Oracle professional" to whom that advice particularly applies.

There are, for example, the inevitable doses of Burlesonian error: "To illustrate, assume I have an index on the last_name column of a 1,000,000 row table and the clustering_factor is 1,000,000 indicating that the rows are in the same sequence as the index". A clustering factor that is equal to the number of rows in the table indicates nothing of the kind, of course.

But whatever: for wimping out, the article can't be beat! "So who is right [those who say you should rebuild indexes or those who say you shouldn't]? I suspect that they are both correct." Uh huh. I suspect that to be both a semantic and logical impossibility myself. But if we wait long enough, I think he might actually just get to the facts of the matter.

We can but hope.

In the context of the present discussion, however, was is most interesting is the distance that apparently now exists between Don and Mike on the issue. I hope the divorce won't be too messy.

Regards
HJR Received on Thu May 06 2004 - 08:25:22 CDT