Subject | Re: [ib-support] Re: Indexs! |
---|---|
Author | Ann W. Harrison |
Post date | 2002-10-01T14:36:46Z |
At 07:19 AM 10/1/2002 +0200, Martijn Tonies wrote:
is a fast index. Notice that for about a quarter million records,
the smaller index was three levels deep while the larger is four.
That means that getting to a specific record takes three reads
in the first case, and four in the second. Since the top level
index node is almost certainly in cache, the difference is actually
2:3.
Regards,
Ann
www.ibphoenix.com
We have answers.
>So Ann, putting the less-selective first is better for avg data lengthActually, it does matter for effectiveness, because a small index
>and doesn't matter for the index effectiveness, right?
is a fast index. Notice that for about a quarter million records,
the smaller index was three levels deep while the larger is four.
That means that getting to a specific record takes three reads
in the first case, and four in the second. Since the top level
index node is almost certainly in cache, the difference is actually
2:3.
Regards,
Ann
www.ibphoenix.com
We have answers.