Re: VIEWS compared to Nodes as Windows into data

<orthogonal>

I'd like to compare views to what, in the world of Oracle, are referred to as "snapshots".

Both views and snapshots use a query, and the data referenced by the query, to produce a result.

The difference is this: the view produces the result on demand, while the snapshot materializes the result in a new table. There is overhead in either case. But in the case of views, the overhead is incurred each time the view is referenced. In the case of snapshots, the overhead is incurred when the snapshot is taken. The snapshot can be further processed or used to construct reports, etc.

What's the disadvantage of snapshots? The data is less current.

VIEWS can be poor performers for a variety of reasons. Among them: sloppy view design, sloppy index design, sloppy table design.

BTW, if we get into the discussion of "good index design" , which is another topic, I suspect that you'll find much more overlap between Oracle expertise and PICK expertise than either camp would like to admit!

</orthogonal>
Received on Thu Apr 29 2004 - 16:07:00 CEST