> > I am not sure about 8.3 but certainly earlier releases of PostgreSQL > > would have specific dependency issues when a sequence was applied to > a > > a column after the fact, versus using the serial or bigserial > > psuedo-types. I'd like to point out that using pg_dump does in fact apply sequences to columns after the fact. (at least in 8.3) Columns lose their "serial" designation after each backup/restore (and therefore during version upgrades) mydb=# create table foo(id serial, bar varchar); NOTICE: CREATE TABLE will create implicit sequence "foo_id_seq" for serial column "foo.id" CREATE TABLE Then, pg_dump produces: -bash-3.00$ pg_dump -s --table=foo mydb CREATE TABLE foo ( id integer NOT NULL, bar character varying ); CREATE SEQUENCE foo_id_seq START WITH 1 INCREMENT BY 1 NO MAXVALUE NO MINVALUE CACHE 1; ALTER SEQUENCE foo_id_seq OWNED BY foo.id; ALTER TABLE foo ALTER COLUMN id SET DEFAULT nextval('foo_id_seq'::regclass); -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general