Re: [HACKERS] Proposal for new SET variables for optimizercosts

Philip Warner wrote:

> Another option would be to add another command, eg. 'PG', which is used for
> all non-SQLxx commands:
> 
>    PG SET somename = somevalue
>    PG VACUUM
> 
> ...etc. But this has the disctinct disadvantage of being more work, and
> being cumbersome in comparison to changing names. 

This does not work out in terms of general SQL compatibility. Even if we
treat commands after PG specially, no other SQL database would, and it
would raise at least as many errors as the extension syntax. Nor is
there any significant advantage of it within Postgres if we ever get a
keyword clash with a future SQL revision - I'd rather not have a syntax
that alows for two interpretations for the same keyword depending on
whether it follows PG or not.

Sevo

-- 
Sevo Stille
sevo@ip23.net