Re: What happens to transactions accessing a view during a REPLACE VIEW?

I would expect this to be a non-starter though it would never occur to me to try it.

But a view being hit thousands of times a second? Why? Why not let them go directly to the tables. What is the advantage of a simple view to a read-only user?

Daniel Morgan

jim rowe wrote:

> Hi,
>
> Suppose i have a simple view pointing to a simple table.
>
> CREATE VIEW my_view AS (select column_a from table_a);
>
> A high performance application is accessing this view thousands of
> times a second. It must run 24 * 7. They are read-only accesses.
>
> I want to replace the view with the following command.
>
> CREATE OR REPLACE VIEW my_view AS (select column_b from table_b).
>
> When i run this command, does the view suffer any 'downtime' during
> the changeover? Will the application be safely blocked from accessing
> the view during the transition, or will some of its accesses to the
> view fail?
>
> What about transactions in progress on the view during the transition.
> Will they be allowed to complete, or will they fail?
>
> Thanks for your help,
> Jim.
Received on Fri Apr 12 2002 - 13:26:59 CDT