Hi,
Jim C. Nasby írta:
On Mon, Jun 12, 2006 at 02:27:31PM +0200, B?sz?rm?nyi Zolt?n wrote:
Zoltan Boszormenyi <[EMAIL PROTECTED]> writes:
after some experimentation, I came up with the attached patch,
which implements parsing the following SERIAL types:
As has been pointed out before, it would be a seriously bad idea to
implement the SQL syntax for identity columns without matching the
SQL semantics for them. That would leave us behind the eight-ball
when we wanted to implement the SQL semantics. Right now we have
a useful but non-standard semantics, and a useful but non-standard
syntax, and those two should stick together.
Well, I read all sections of 5WD-02-Foundation-2003-09.pdf
where "identity" appears, here are the list of changes that will
be needed for an identity column:
Have you read the archives on the recent discussions that have taken
place about whether SERIAL should be a black box or not? IIRC most of
this was all hashed out in that thread.
I just read it thoroughly, and the issues I listed wasn't mentioned
in the "black box" thread, at all. I am trying to implement the
standard syntax ( and gradually the conformant behaviour )
along the lines of sections 4.14.7, 11.3, 11.4, 11.7, 11.11,
11.12, 11.17 and 14.8.
Best regards,
Zoltán Böszörményi
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match
