在Postgresql 9.1中使用动态表名进行循环？

问题描述

我有一个 plpgslq 函数，该函数进行一些数据处理，并想编写一个 for 循环，但是在设计时未知表名。有什么可能的方法来实现这一目标？这是我想要实现的示例代码片段：

I have a plpgslq function which does some data processing and would like to write a for loop, however my table name is not known at design time. Is there any possible way to achieve this? Here is sample code snippet of what I want to achieve:

-- Function: check_data()

-- DROP FUNCTION check_data();

CREATE OR REPLACE FUNCTION check_data()
  RETURNS character varying AS
$BODY$declare
 dyn_rec record;
 tbl_name record;
begin
  -- sample dynamic tables
  tbl_name := 'cars';
  tbl_name := 'trucks';
  tbl_name := 'bicycles';

  for dyn_rec in select * from format($$s%$$,tbl_name) loop
    raise notice 'item is %',dyn_rec.item_no;
  end loop;

  return 'Processing Ok';

end;$BODY$
  LANGUAGE plpgsql VOLATILE
  COST 100;
ALTER FUNCTION check_data()
  OWNER TO postgres;

推荐答案

您不能将变量用作表或列的标识符plpgsql曾经有过嵌入式SQL。解决方案是动态SQL-EXECUTE或FOR IN EXECUTE语句：

You cannot use a variable as table or column identifier in plpgsql embedded SQL ever. A solution is dynamic SQL - EXECUTE or FOR IN EXECUTE statements:

DO $$
DECLARE
  tables text[] = ARRAY['table1','table2'];
  table_name text;
  rec record;
BEGIN
  FOREACH table_name IN ARRAY tables
  LOOP
    FOR r IN EXECUTE format('SELECT * FROM %I', table_name)
    LOOP
      RAISE NOTICE '%', rec;
    END LOOP; 
  END LOOP;
END; $$

这篇关于在Postgresql 9.1中使用动态表名进行循环？的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！