根据postgres sql函数中的SELECT结果有条件地执行INSERT

Question

如何根据另一个表是否有值来对表执行插入操作？下面是一个例子

create table mygroup (
  group_id integer primary key,
  other_id integer not null references othertable(other_id)
);

create table myitem (
  group_id integer not null references mygroup(group_id),
  item_id integer not null
);

create function add_to_group (group_arg integer, item_arg integer, other_arg integer) language sql as $$
  select * from mygroup where group_id = group_arg and other_id = other_arg;
  -- Do the next statement only if the previous one had a result
  insert into myitem (group_id, item_id) values (group_arg, item_arg);
$$;

如果我使用的是plpgsql函数，就可以用if (found)来实现。但是我如何使用普通的sql函数来实现它呢？有没有办法将这两条语句合并为一条语句，比如使用JOIN执行INSERT？

Answer 1

我想你只是想要exists

insert into myitem (group_id, item_id) 
    select v.group_id, v.item_id
    from (values (group_arg, item_arg)) v(group_id, item_id)
    where exists (select 1
                  from mygroup g
                  where g.group_id = v.group_id and v.other_id = other_arg
                 );

或者，如果mygroup表中只有一行匹配，则可以使用select：

insert into myitem (group_id, item_id) 
    select group_arg, item_arg
    from mygroup g
    where g.group_id = group_arg and v.other_id = other_arg;

如果可能存在重复项，您可以将其调整为：

insert into myitem (group_id, item_id) 
    select distinct group_arg, item_arg
    from mygroup g
    where g.group_id = group_arg and g.other_id = other_arg;