Re: "SQL" REPLACE SYNTAX - Mailing list pgsql-php

From Manuel Lemos
Subject Re: "SQL" REPLACE SYNTAX
Date
Msg-id 3C4E5013.CA5C24FC@acm.org
Whole thread Raw
In response to Re: "SQL" REPLACE SYNTAX  ("Christopher Kings-Lynne" <chriskl@familyhealth.com.au>)
List pgsql-php
Hello,

Christopher Kings-Lynne wrote:
>
> > Look at Metabase implementation. Metabase comes with a driver
> > conformance test that verifies if features like this work properly in
> > all supported databases. Don't you think that I would release a version
> > of Metabase that would not pass that test if all databases?
> >
> > The application needs to hint it as it is something that does not change
> > at runtime, so there is no penalty of figuring that by querying the
> > database server and some database do not provide a way to tell you which
> > fields are the primary key.
>
> OK, so it's not a simulation of the MySQL syntax - MySQL requires no hints.

No, but the functionality is the same. MySQL ignores the hints because
SQL REPLACE works as intended without the hints.



> I've looked thru the code and yeah, basically that's what you have to do, so
> long as you can get the app programmer to tell you in the code what the
> primary keys are.  For a postgres implementation, it's more difficult.

It is required. If the programmer does not tell which fields are the
keys, the function will fail with an error stating that.



> BTW, there's no documentation anywhere in your code - makes it a bit
> difficult to see what's going on...

Erm.... there is 248K HTML document that details all about that function
the whole Metabase API. :-)

Regards,
Manuel Lemos

pgsql-php by date:

Previous
From: Manuel Lemos
Date:
Subject: Re: "SQL" REPLACE SYNTAX
Next
From: Faudzy Sulaiman
Date:
Subject: unable to connect to PostgreSQL server