虽然真的很鸡肋,并且性能不杂滴,但聊胜于无吧.

drop function if exists exe_dynamic_sql(bigint);
drop function if exists exe_dynamic_count(bigint);
--返回记录集
create or replace function exe_dynamic_sql(ival bigint)
    returns table(objectid bigint,name varchar(128))
as $$
	declare
	begin
		return query execute 'select objectid,name from dictionarys where parentid=$1 order by parentid,sort' using $1; 
	end;
$$ language plpgsql;

--赋值给变量
create or replace function exe_dynamic_count(ival bigint)
    returns bigint
as $$
	declare
		v_count bigint;
	begin
		execute 'select count(*) from dictionarys where parentid=$1' using $1 into v_count;
		return v_count;
	end;
$$ language plpgsql;
--测试
select * from exe_dynamic_sql(26);
select exe_dynamic_count(26);

要点:

• returns table(objectid bigint,name varchar(128)),定义返回的字段和类型
• using $1执行时使用过程参数;

准备一个语句用于执行,这个就比较重要了,查询参数绑定,开发利器

有朋友抱怨同一sql时快时慢,这是因为值在表中的占比不同,占比小的值就可以使用索引,值占比超过5%里sql就很慢了.此时就可以用下面的sql调式sql,可以根据不同的值来观察执行计划.

--getDictionarys仅在当前会话下有效
prepare getDictionarys (bigint) as
    select objectid,name from dictionarys where parentid=$1 order by parentid,sort;

explain (analyze,verbose,costs,buffers,timing)
execute getDictionarys(24);

explain (analyze,verbose,costs,buffers,timing)
execute getDictionarys(25);

explain (analyze,verbose,costs,buffers,timing)
execute getDictionarys(26);

--释放指定的预备语句
deallocate getDictionarys;
--释放所有预备语句
deallocate all;