home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.databases.theory
- Path: sparky!uunet!enterpoop.mit.edu!bloom-picayune.mit.edu!thomasl
- From: thomasl@mtl.mit.edu (Thomas J Lohman)
- Subject: Re: Materialised views
- Message-ID: <1993Jan7.185243.2625@athena.mit.edu>
- Sender: news@athena.mit.edu (News system)
- Nntp-Posting-Host: garcon.mit.edu
- Organization: MIT Microsystems Technology Laboratories
- References: <C0HtzK.BEL@gabriel.keele.ac.uk>
- Date: Thu, 7 Jan 1993 18:52:43 GMT
- Lines: 23
-
- In article <C0HtzK.BEL@gabriel.keele.ac.uk> csa09@keele.ac.uk (Paul Singleton) writes:
-
- A note on "materialised views". I think things would get very messy
- when the underlying db table structure changed for any reason or when data
- is added to the underlying tables. The db manager would have to
- make checks to see if any of the newly added data affects the stored
- view data.
-
- >Also, do any commercial database products provide special storage or
- >indexing schemes for unchanging data? I.e. which do not compromise
- >towards the need to incrementally change the data, as B-trees etc. do.
-
- Yes, some commercial database products, such as Ingres, give you the
- ability to remodify your tables to a static version of B-tree. They also
- give you the ability to reorganize it based on hashed key, which they say
- gives you the quickest access for exact key matches. If the table is
- unchanging (static) then these would be your best storage structures.
-
-
- --tom
-
-
-
-