开发者

Updating the summary table based on triggers and stored procedures

开发者 https://www.devze.com 2023-04-08 13:28 出处:网络
I have a typical LAMP based site + Zend Framework where I have a base table and a summary table. Summary table is used to display data in reports.

I have a typical LAMP based site + Zend Framework where I have a base table and a summary table. Summary table is used to display data in reports.

Base table -

I开发者_开发问答D | Status
1 | 1
2 | 1
3 | 2
4 | 2
5 | 1
6 | 1

Summary table -

Status | Count
1 | 4
2 | 2

The base table will be changed(insert,update,delete) at an average of 20 times per day.

Currently, I am using triggers to call a stored procedure which will update the summary table based on the base table.

This is the stored procedure.

CREATE PROCEDURE UpdateSummary()

BEGIN

UPDATE summary a

INNER JOIN

(SELECT status, count(*) c from base group by status) b

ON a.status = b.status

SET a.count = b.c;

END

And I have 3 triggers (one for each - Insert, Delete and Update). I have shown the insert sample alone below. Other are similar to this.

CREATE TRIGGER S_T_TRIGGER_I

AFTER INSERT ON base

FOR EACH ROW

CALL UpdateSummary();

I want the summary table to be updated to the latest values always. Using triggers and stored procedure like this is the best way or is there a elegant way to do this?


Well you are re-querying the DB over and over for data that you already know.

Why not just update the summary with only the changes.

DELIMITER $$

CREATE TRIGGER ai_base_each AFTER INSERT ON base FOR EACH ROW
BEGIN
  INSERT INTO summary (status, count) VALUES (NEW.status,1)
  ON DUPLICATE KEY UPDATE
    SET count = count + 1;
END $$

CREATE TRIGGER ad_base_each AFTER DELETE ON base FOR EACH ROW
BEGIN
  UPDATE summary s
    SET s.count = s.count - 1 
    WHERE s.status = OLD.status;
END $$


CREATE TRIGGER au_base_each AFTER UPDATE ON base FOR EACH ROW
BEGIN
  UPDATE summary s
    SET s.count = s.count - 1 
    WHERE s.status = OLD.status;

  INSERT INTO summary (status, count) VALUES (NEW.status,1)
  ON DUPLICATE KEY UPDATE
    SET count = count + 1;
END $$

DELIMITER ;

This will be much much faster and more to the point much more elegant.


Why don't you use a view like :

 CREATE VIEW Summary AS 
      SELECT status, count(*)
      FROM Base
      GROUP BY status;

Each time you need, just do :

SELECT * 
FROM Summary 

And you'll get your result in real time (each call re-computed).

Views can be used the same way like table is used in Zend Framework. Just that you need to specify a primary key explicitly as explained here

0

精彩评论

暂无评论...
验证码 换一张
取 消