MySQL中distinct和group by去重性能對比

前言

• MySQL：5.7.17
• 存儲引擎：InnoDB
• 實驗目的：本文主要測試在某字段有無索引、各種不同值個數情況下，記錄對此字段其使用DISTINCT/GROUP BY去重的查詢語句執行時間，對比兩者在不同場景下的去重性能，實驗過程中關閉MySQL查詢緩存。
• 實驗表格：

表名	記錄數	查詢字段有無索引	查詢字段不同值個數	DISTINCT	GROUP BY
tab_1	100000	N	3
tab_2	100000	Y	3
tab_3	100000	N	10000
tab_4	100000	Y	10000

---

實驗過程

1）創建測試表

表創建語句：

DROP TABLE IF EXISTS `tab_1`;
CREATE TABLE `tab_1` (
  `id` int(10) unsigned NOT NULL AUTO_INCREMENT,
  `value` int(10) unsigned NOT NULL,
  PRIMARY KEY (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8;

DROP TABLE IF EXISTS `tab_2`;
CREATE TABLE `tab_2` (
  `id` int(10) unsigned NOT NULL AUTO_INCREMENT,
  `value` int(10) unsigned NOT NULL,
  PRIMARY KEY (`id`),
  KEY `idx_value` (`value`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8;

DROP TABLE IF EXISTS `tab_3`;
CREATE TABLE `tab_3` LIKE `tab_1`;

DROP TABLE IF EXISTS `tab_4`;
CREATE TABLE `tab_4` LIKE `tab_2`;

2）生成測試數據

表數據插入過程：

DROP PROCEDURE IF EXISTS generateRandomData;
delimiter $$
-- tblName爲插入表,field爲插入字段,num爲插入字段值上限,count爲插入的記錄數
CREATE PROCEDURE generateRandomData(IN tblName VARCHAR(30),IN field VARCHAR(30),IN num INT UNSIGNED,IN count INT UNSIGNED)
BEGIN
	-- 聲明循環變量
	DECLARE i INT UNSIGNED DEFAULT 1;
	-- 循環插入隨機整數1~num,共插入count條數據
	w1:WHILE i<=count DO
		set i=i+1;
		set @val = FLOOR(RAND()*num+1);
		set @statement = CONCAT('INSERT INTO ',tblName,'(`',field,'`) VALUES(',@val,')');
		PREPARE stmt FROM @statement;
		EXECUTE stmt;
	END WHILE w1;
END $$
delimiter ;

調用過程隨機生成測試數據：

call generateRandomData('tab_1','value',3,100000);
INSERT INTO tab_2 SELECT * FROM tab_1;

call generateRandomData('tab_3','value',10000,100000);
INSERT INTO tab_4 SELECT * FROM tab_3;

3）執行查詢語句，記錄執行時間

查詢語句及對應執行時間如下：

SELECT DISTINCT(`value`) FROM tab_1;
SELECT `value` FROM tab_1 GROUP BY `value`;

SELECT DISTINCT(`value`) FROM tab_2;
SELECT `value` FROM tab_2 GROUP BY `value`;

SELECT DISTINCT(`value`) FROM tab_3;
SELECT `value` FROM tab_3 GROUP BY `value`;

SELECT DISTINCT(`value`) FROM tab_4;
SELECT `value` FROM tab_4 GROUP BY `value`;

4）實驗結果

表名	記錄數	查詢字段有無索引	查詢字段不同值個數	DISTINCT	GROUP BY
tab_1	100000	N	3	0.058s	0.059s
tab_2	100000	Y	3	0.030s	0.027s
tab_3	100000	N	10000	0.072s	0.073s
tab_4	100000	Y	10000	0.047s	0.049s

---

實驗結論

MySQL 5.7.17中使用distinct和group by進行去重時，性能相差不大

實驗過程及結論，如有不足之處，歡迎指正，此實驗結論僅供參考。

---

End~