Question

有我的桌子：

create table location (
    nom text primary key,
    adresse text
);
create table groupe (nom text primary key);
create table groupelocation (
    nomGroupe text references groupe (nom) on delete cascade,
    nomLocation text references location (nom) on delete cascade,
    primary key(nomGroupe, nomLocation)
);
insert into groupe values('groupe');
insert into location values('location', 'là bas');
insert into groupelocation values('groupe', 'location');

我希望当我删除组表中的somethng时，删除所有关联的位置，但我找不到解决方案：/ 我试过这样做：create rule deletelocations as on delete to groupe do instead (delete from (select location natural join groupelocation where old.nom = nomGroupe));但这不起作用：/你能帮我吗？

Answer 1

Postgres不支持将CASCADE提供给DELETE语句，但它支持在表创建时级联删除。

http://www.postgresql.org/docs/9.0/static/sql-createtable.html

CASCADE删除引用已删除行的所有行，或更新引用列的值为引用的新值专栏。

Answer 2

有一个解决方案，我已经创建了一个小的plpgsql函数来解决这个问题

CREATE OR REPLACE FUNCTION deletelocation()
RETURNS text AS $$
DECLARE
    idloc int;
    curs refcursor;
BEGIN
    OPEN curs FOR SELECT location.id FROM location except SELECT location.id FROM location JOIN groupelocation ON location.id = groupelocation.idLocation;
    LOOP
        FETCH curs INTO idloc;
        EXIT WHEN NOT FOUND;
        DELETE FROM location WHERE id = idloc;
    END LOOP;
    CLOSE curs; 
    RETURN '';
END;
$$ language plpgsql;

PGSQL删除级联

2 个答案: