Re: The "Why Postgres not MySQL" bullet list

On Sun, 2003-08-31 at 16:21, Ian Barwick wrote:
> On Sunday 31 August 2003 21:32, Rod Taylor wrote:
> > Very interesting read -- and I'm only through section 1. Having used
> > MySQL for a while some are known to me, many others are new.
> >
> > One particularly interesting item is Section 1.7, DEFAULT 'now()' as
> > this is also a PostgreSQL gotcha. See 'nowtest' attached.
>
> Interesting. Though I think this is subtly different - in MySQL the default is
> ignored, you could put anything in you want...
>
> > Anyway, I noticed a few times the spec has been mentioned as being
> > broken, ignored or otherwise, but the relevant sections not pointed to.
> >
> > Would you like a hand in tracing down quotes where spec in these areas?
>
> Yup. I'm interested in facts not FUD ;-).
>
> On that note I do have vague intentions of compiling a list
> of PostgreSQL gotchas. The only other one I have right now is -

Implicit from items (toggle in 7.4, but allows them by default).

Constraint names are unique to a table rather than a schema.

Folds identities to lowercase rather than uppercase when they are not
quoted.

Unique constraints are evaluated immediately. This can make UPDATE ..
SET col = col + 1 evaluation difficult.

That is all I have for PostgreSQL at the moment.

For Gotcha 1.8 (INSERT INTO .. SELECT)

Self referencing operations are an advanced feature which MySQL does not
support.

Conformance Rules (Section 14.8)

1) Without Feature F781, Self-referencing operations, no leaf
generally underlying table of T shall be generally contained in
the <query expression> immediately contained in the <insert
columns and source> except as the <table or query name> or
<correlation name> of a column reference.