At 05:14 PM 6/9/2001 -0400, Claudio Valderrama C. wrote:

> While the question was targeted to Jason, I like to point some facts:
>- Probably the need to drop a table is to create a new one with the same
>name but different structure. Most common case?

That's what I thought. The syntax for RECREATE TABLE is the same as
the syntax for CREATE TABLE. The difference is that RECREATE will
delete the previous version of the table if one exists.

>- It may be the case also that you want to drop and define it same as before
>just because it may be more efficient than deleting all records in the
>table. Here I know about a weird latency problem, where at least two people
>reported last year that recreating the table as fast as possible (doing drop
>and create in an script) caused the new table to have records.

The "timing" shouldn't be a problem, though there's a chance that dropping
and creating in the same transaction is a problem.

>- There may be some tables with fixed structure that are used as helpers for
>some complex operations. Maybe the script needs to simply drop them without
>caring if they exist or not.

That's what I wondered...

>- The idea of morphing a table to create a new one is nice in theory. If it
>works in NetfraS, good for it. However, at this time, in FB it's unfeasible
>IMHO: the mosquito (pardon, the bird) has today enough trouble altering
>metadata to introduce yet another headache.

I'm not proposing that RECREATE morph the data - which Jim's UPDATE may
do - just delete the table if it exists and define a new table.


Regards,

Ann
www.ibphoenix.com
We have answers.