>I understand the ease of converting to SQL and the ease of using views. But using views on a local data set has caused performance problems. The client does not want to purge data (we are talk about only two years) and the thought of installing SQL (anybody's) is a little much when difference is how the coding is done. Rememeber using tableupdate(1) on 1.3 million records is causing long delays and using replace/insert is working within a few seconds. How can I justify the use of views?
>
>John

Hey John,

I can appreciate how you'd have performance problems with local views in a table of 1.3 million records. Just a personal observation:

1. With tables that large I'd push the client *real hard* to upsize to Oracle. Keep in mind that in the scheme of RDBMS SQL Server falls into the *workgroup* category and Oracle (or Sybase and a few others) fall into the *enterprise category.

2. If you're coded for local views your are 90% coded for remote views. The only thing you need change are some of your SELECT statments, so I think you'd find if you do it that the difference in coding is really not that much.

Cheers.