interesting! a sequence clashes with data already in that table

[Miles Keaton]

Here's an interesting problem!

When a sequence clashes with data already in that table:

CREATE TABLE clients (
id serial NOT NULL PRIMARY KEY UNIQUE,
name varchar(64));

-- import OLD clients, with their original ID#...
INSERT INTO clients VALUES (3, 'Dave');
INSERT INTO clients VALUES (4, 'Sally');

-- done! let the world add new clients now

INSERT INTO clients (name) VALUES ('Harry');
-- no problems, id=1
INSERT INTO clients (name) VALUES ('Mary');
-- no problems, id=2
INSERT INTO clients (name) VALUES ('Burt');
ERROR: duplicate key violates unique constraint "clients_pk ey"
I thought Postgres would be smart enough to make the clients_id_seq
skip existing numbers, but I guess not!

Do I, instead, have to be smart enough to set the sequence over the
highest existing id# from my imported data, so that it can blindly
increment without clashing?

If so, setval() would be best for that, right?

---------------------------(end of broadcast)---------------------------
TIP 4: Don't 'kill -9' the postmaster

[Bruno Wolff III]

On Fri, Oct 08, 2004 at 03:00:07 -0700,
Miles Keaton <mi*********@gm ail.com> wrote:

I thought Postgres would be smart enough to make the clients_id_seq
skip existing numbers, but I guess not!
It wouldn't be very efficient to do that.
Do I, instead, have to be smart enough to set the sequence over the
highest existing id# from my imported data, so that it can blindly
increment without clashing?
That is normally what you want to do.
If so, setval() would be best for that, right?

Yes.

---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match