Hi,
I have a question about the UNIQUE constraint. The documentation
describes this example:
CREATE TABLE example (
a integer,
b integer,
c integer,
UNIQUE (a, c)
);
But it is not clean to me. Does the above example mean that the list
of pairs must be unique or is it only a
I have a table with fields:
id A B C D E F
where id is an int4 primary key.
In this table there is information like:
1 a1 b1 xxx xxx
2 a1 b1 xxx xxx xxx xxx
3 a2 b2 xxx
I create every time i need this, a copy of this field filled out by a
trigger
on insert and update that holds the upper (or lower for you) value of
the orginal field like this:
create table users (
email varchar(255),
...
u_email varchar(255)
...
);
CREATE OR REPLACE FUNCTION
On Aug 7, 2004, at 3:25 AM, Sascha Ziemann wrote:
CREATE TABLE example (
a integer,
b integer,
c integer,
UNIQUE (a, c)
);
But it is not clean to me. Does the above example mean that the list
of pairs must be unique
Yes.
Does the following table fullfill the UNIQUE clause of the
Kenneth,
but why would anyone want to change the value of an autogenerated serial
row?
But if you're using a real key, it may need to change. The only reason *not*
do do it that way is performance issues with CASCADE.
--
-Josh Berkus
Aglio Database Solutions
San Francisco
