Re: use of Mixed DB types.

From: Carl Rosenberger <carl_at_db4o.com>
Date: Tue, 20 Nov 2001 11:49:17 +0100
Message-ID: <9tdcdj$jr0$04$1_at_news.t-online.com>


Kendall wrote:
> table is changed to class
> view is changed to virtual class
> row of table is changed to instance of class
> column is changed to attribute
> procedure is changed to method
> table hierarchy is changed to class hierarchy
> child table is changed to subclass
> parent table is changed to superclass

...
[snip]
...

> Most of the major RDBMS's now have
> object support in them as well.

Most of the "major RDBMS" treat objects differently than you have described it here:
"Object maps to column of row"

I wouldn't call that "Object support", since it lacks querying functionality and it typically is a nightmare to change class definitions.

Let's call that "structurized blob support". You can also achieve that by serializing objects to byte arrays.

Object databases provide more functionality: - They manage object identities and will ensure that objects are instantiated only once, no matter by how many queries the object is accessed at the same time.
- They provide navigation features to walk through an object graph by instantiating objects at the users requests without the need for repetitive queries. - ID lookups typically are extremely fast. - They allow refactoring classes within existing database files.
- Some provide automatic schema management. You don't need to manage the class model, since the database can analyze your classes.
- ...and more...

Kind regards,
Carl

---
Carl Rosenberger
db4o - database for objects - http://www.db4o.com
Received on Tue Nov 20 2001 - 11:49:17 CET

Original text of this message