示例表：

CREATE TABLE students
(
  sno character varying(32) NOT NULL,--学号
  sname character varying(64) NOT NULL,--姓名
  sage integer,--年龄
  scid character varying(32),-- 班级id
  scname character varying(64),-- 班级名称
  CONSTRAINT students_pkey PRIMARY KEY (sno)
)

这里的需求是查出班级以及这个班里的学生，查询函数如下：

CREATE OR REPLACE FUNCTION getstudents(
    OUT cid character varying,OUT cnm character varying,OUT sts character varying[])
  RETURNS SETOF record AS
$BODY$
	declare
		sc record;
	begin
		for sc in (select distinct(scid),scname from students) loop
			cid := sc.scid;
			cnm := sc.scname;
			select array(select sno||' '||sname||' '||sage from students where scid = sc.scid order by sage) into sts;
			return next;
		end loop;
	end;

$BODY$
  LANGUAGE plpgsql VOLATILE
  COST 100
  ROWS 1000;
ALTER FUNCTION getstudents()
  OWNER TO postgres;

select * from getStudents();

语法糖衣

以上是比较简单传统的方式，还有更加简单的方式，如下：

select scid,scname,array_agg(sno||' '||sname||' '||sage)as stus from students group by scid,scname;

array_agg是特殊的聚集函数，可以将属于同一个的group的记录放到一个数组中。很好用！