MySQL 大致测试更新时间

1：需求：把一个2千万条数据的一个表，随机更新其中的二十行需要大致多久？

DROP TABLE IF EXISTS test20;

　　

CREATE TABLE test20(
id INT AUTO_INCREMENT PRIMARY KEY,
money DOUBLE,
createdtime DATETIME)ENGINE=INNODB DEFAULT CHARSET=UTF8;

　　

DROP PROCEDURE IF EXISTS pro_test20;
 
DELIMITER &&
CREATE PROCEDURE pro_test20(IN inr int)
  BEGIN
  DECLARE i INT DEFAULT 0;
  START TRANSACTION;
   WHILE i<inr DO
    INSERT INTO test20(money,createdtime) values(234567891023,'2014-06-22 20:18:38');
    SET i=i+1;
   END WHILE;
  COMMIT;
  END && 
 
DELIMITER ;

mysql> call pro_test20(10000000);
Query OK, 0 rows affected (4 min 50.37 sec)

　　

mysql> call pro_test20(10000000);
Query OK, 0 rows affected (8 min 18.97 sec)

　　

#20次
INSERT INTO test20(money,createdtime) values(234567891023,'2013-03-33 20:18:38');

mysql> INSERT INTO test20(money,createdtime) values(234567891023,'2013-03-03 20:18:38');
Query OK, 1 row affected, 1 warning (0.02 sec)

UPDATE test20 SET money=money+10000000,createdtime=now()
WHERE createdtime='2013-03-03 20:18:38';

mysql> UPDATE test20 SET money=money+10000000,createdtime=now() WHERE id BETWEEN 200000 AND 200020;
Query OK, 21 rows affected (0.01 sec)
Rows matched: 21  Changed: 21  Warnings: 0

mysql> UPDATE test20 SET money=money+10000000,createdtime=now() where id in (36,457,67842,21,324,68,23,12,345,547,56,234,7263,233525,252522,3523,76666,1231,54);
Query OK, 19 rows affected (0.05 sec)
Rows matched: 19  Changed: 19  Warnings: 0

　　

　　

1：需求：把一个2千万条数据的一个表，随机更新其中的一千行需要大致多久？

存储过程测试1：（效率很低）

DROP PROCEDURE IF EXISTS pro_tab1000;
DROP TABLE IF EXISTS tab_test20;
CREATE TABLE tab_test20(id int)ENGINE=INNODB DEFAULT CHARSET=UTF8;

DELIMITER &&
CREATE PROCEDURE pro_tab1000(IN inr int)
  BEGIN
  DECLARE i INT DEFAULT 0;
   WHILE i<inr DO
    INSERT INTO tab_test20 SELECT floor(RAND() * (SELECT MAX(id) FROM test20));
    SET i=i+1;
   END WHILE;
  END && 
 
DELIMITER ;

CALL pro_tab1000(1000);


UPDATE test20 SET money=money+10000000,createdtime=now() WHERE id in (SELECT DISTINCT id FROM tab_test20);

　　

存储过程测试2：2s

DELIMITER &&
DROP PROCEDURE IF EXISTS pro_test1000;
CREATE PROCEDURE pro_test1000(IN inr int)
  BEGIN
  DECLARE i INT DEFAULT 0;
  DECLARE j INT ;
  SELECT FLOOR(RAND()*(SELECT MAX(id) FROM test20)) into j;
   WHILE i<inr DO
    UPDATE test20 SET money=money+10000000,createdtime=now() WHERE id = j;
    SET i=i+1;
   END WHILE;
  END && 
 
DELIMITER ;

　　

效率很低的语句：

SELECT id FROM test20 ORDER BY rand() LIMIT 10

　　

mysql> SELECT id FROM test20 ORDER BY rand() LIMIT 10;
+----------+
| id       |
+----------+
|  1876355 |
| 10266755 |
| 14746331 |
| 17533398 |
|  8164806 |
|  8438406 |
| 12884382 |
| 17285257 |
|  9817314 |
| 10273314 |
+----------+
10 rows in set (26.67 sec)

　　

问题1：更新记录很多的时候，字段上有索引快还是没有索引快？