home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!gossip.pyramid.com!pyramid!infmx!davek
- From: davek@informix.com (David Kosenko)
- Newsgroups: comp.databases.informix
- Subject: Re: Estimating sizes of indexes
- Message-ID: <1992Nov11.151731.7281@informix.com>
- Date: 11 Nov 92 15:17:31 GMT
- References: <2AFFE3CA.81BE@tct.com>
- Sender: news@informix.com (Usenet News)
- Organization: Informix Software, Inc.
- Lines: 26
-
- Jim Kunzman writes:
- >Informix provides very useful tools for estimating sizes of rows in
- >a table. Perhaps I've overlooked it, but I can find nothing about
- >estimating sizes for indexes. In other words, given table X with
- >indexes A, B, and C, how much overhead is required for each additional
- >row in table X?
- >
- >Can this be done easily? Is there somewhere in the documentation where
- >this is covered? Is there any difference in overhead for SE and Online
- >(unfortunately, we have to support both)?
-
- For OnLine, this is covered (in the Guide to SQL Tutorial, pg 10-16) under
- Estimating Index Pages. There is some difference between SE and OnLine -
- to keep it simple, consider SE's "page size" to be 1K (since index "nodes"
- are 1 page in OnLine and 1K in SE). The resulting number is really an
- approximation, since it is impossible to accurately predict b+tree
- overhead without knowing how and when the keys were inserted, exactly how
- many duplicates of each key (when applicable) etc., but it can get you
- pretty much in the ballpark.
-
- Dave
- --
- Disclaimer: These opinions are not those of Informix Software, Inc.
- **************************************************************************
- "I look back with some satisfaction on what an idiot I was when I was 25,
- but when I do that, I'm assuming I'm no longer an idiot." - Andy Rooney
-