At 10:20 PM 10/12/2008, you wrote:

> This is, in fact, about displaying a progress window.
>
> I start a transaction, then get the count of records to
>be read, and display that. Then I run the select, updating
>the display of records read after every 100 fetches, and
>then I commit the transaction.
>
> This application is currently using the embedded server,
>so at this time the database won't be changed. And if it
>is, so what? We have to start a transaction at some point,
>and if the
>
> So I am using this count for exactly the application you
>say is inefficient. How would you do this?

I wouldn't write an application that was pulling over so many records that it needed a progress window to entertain the hapless waiting user. Sorry.

./hb