Hi all, I have a scenario like follow (script generated with pg_dump) :
CREATE TABLE teste ( id integer NOT NULL, ds character varying(20), n2 character varying(10) ); CREATE SEQUENCE teste_id_seq INCREMENT BY 1 NO MAXVALUE NO MINVALUE CACHE 1; ALTER TABLE teste ALTER COLUMN id SET DEFAULT nextval('teste_id_seq'::regclass); ALTER TABLE ONLY teste ADD CONSTRAINT pk_teste PRIMARY KEY (id); ALTER SEQUENCE teste_id_seq OWNED BY teste.id; CREATE VIEW vteste AS SELECT teste.id, teste.ds, teste.n2 FROM teste; ALTER TABLE vteste ALTER COLUMN id SET DEFAULT nextval('teste_id_seq'::regclass); Well this last line make a change on a view column and it works. I do it because when inserting a row in the view vteste, I got a message error like "teste.id does not accept NULL values" Searching in the internet, I found a message where Tom Lane suggest to do this. Well, I don´t found anything in Postgresql manual saying that we can do it in views, but I don´t found anything saying that we don´t must do it too. Since pg_dump generates the script with this definition, I think that they must be supported by PostgreSQL The only thing that I have a objection is because this don´t appear in pgadmin views definition. And we don´t mind that this "ALTER COLUMN" exists in the Database structure. I want to request/suggest that views script generates a complete script in the sqlpane like when select a table (with index,triggers creation sql scripts) or with at least this DDL, since we can´t see it anywhere. May we have a icon for this alter columns in the view definition tree too ? I´m using pgAdmin 1.6.3 in a winXP pro SP2 with Postgresql 8.2.4 Thanks in Advance Luiz ---------------------------(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