MySQL quickly creates tens of millions of test data

Note: The data volume of this article is 100W. If you want tens of millions, just increase the amount. However, do not use rand() or uuid() in large quantities, which will cause performance degradation.

Background

When performing performance testing of query operations or sql optimization, we often need to build a large amount of basic data in the offline environment for our testing to simulate the real online environment.

Nonsense, you can’t let me test online, I will be hacked to death by the DBA

How to create test data

    1. 编写代码，通过代码批量插库（本人使用过，步骤太繁琐，性能不高，不推荐）
    2. 编写存储过程和函数执行（本文实现方式1）
    3. 临时数据表方式执行 （本文实现方式2，强烈推荐该方式，非常简单，数据插入快速，100W，只需几秒）
    4. 一行一行手动插入，（WTF，去死吧）

Create the basis Table structure

No matter which method is used, the table I want to insert must be created

CREATE TABLE `t_user` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `c_user_id` varchar(36) NOT NULL DEFAULT '',
  `c_name` varchar(22) NOT NULL DEFAULT '',
  `c_province_id` int(11) NOT NULL,
  `c_city_id` int(11) NOT NULL,
  `create_time` datetime NOT NULL,
  PRIMARY KEY (`id`),
  KEY `idx_user_id` (`c_user_id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4;

Method 1: Using stored procedures and memory tables

• Create memory table

利用 MySQL 内存表插入速度快的特点，我们先利用函数和存储过程在内存表中生成数据，然后再从内存表插入普通表中

CREATE TABLE `t_user_memory` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `c_user_id` varchar(36) NOT NULL DEFAULT '',
  `c_name` varchar(22) NOT NULL DEFAULT '',
  `c_province_id` int(11) NOT NULL,
  `c_city_id` int(11) NOT NULL,
  `create_time` datetime NOT NULL,
  PRIMARY KEY (`id`),
  KEY `idx_user_id` (`c_user_id`)
) ENGINE=MEMORY DEFAULT CHARSET=utf8mb4;

• Create functions and stored procedures

# 创建随机字符串和随机时间的函数
mysql> delimiter $$
mysql> CREATE DEFINER=`root`@`%` FUNCTION `randStr`(n INT) RETURNS varchar(255) CHARSET utf8mb4
    ->     DETERMINISTIC
    -> BEGIN
    ->     DECLARE chars_str varchar(100) DEFAULT 'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789';
    ->     DECLARE return_str varchar(255) DEFAULT '' ;
    ->     DECLARE i INT DEFAULT 0;
    ->     WHILE i          SET return_str = concat(return_str, substring(chars_str, FLOOR(1 + RAND() * 62), 1));
    ->         SET i = i + 1;
    ->     END WHILE;
    ->     RETURN return_str;
    -> END$$
Query OK, 0 rows affected (0.00 sec)

mysql> CREATE DEFINER=`root`@`%` FUNCTION `randDataTime`(sd DATETIME,ed DATETIME) RETURNS datetime
    ->     DETERMINISTIC
    -> BEGIN
    ->     DECLARE sub INT DEFAULT 0;
    ->     DECLARE ret DATETIME;
    ->     SET sub = ABS(UNIX_TIMESTAMP(ed)-UNIX_TIMESTAMP(sd));
    ->     SET ret = DATE_ADD(sd,INTERVAL FLOOR(1+RAND()*(sub-1)) SECOND);
    ->     RETURN ret;
    -> END $$

mysql> delimiter ;

# 创建插入数据存储过程
mysql> CREATE DEFINER=`root`@`%` PROCEDURE `add_t_user_memory`(IN n int)
    -> BEGIN
    ->     DECLARE i INT DEFAULT 1;
    ->     WHILE (i          INSERT INTO t_user_memory (c_user_id, c_name, c_province_id,c_city_id, create_time) VALUES (uuid(), randStr(20), FLOOR(RAND() * 1000), FLOOR(RAND() * 100), NOW());
    ->         SET i = i + 1;
    ->     END WHILE;
    -> END
    -> $$
Query OK, 0 rows affected (0.01 sec)

• Call stored procedure

mysql> CALL add_t_user_memory(1000000);
ERROR 1114 (HY000): The table 't_user_memory' is full
出现内存已满时，修改 max_heap_table_size 参数的大小，我使用64M内存，插入了22W数据，看情况改，不过这个值不要太大，默认32M或者64M就好，生产环境不要乱尝试

• Insert into ordinary table from memory table

mysql> INSERT INTO t_user SELECT * FROM t_user_memory;
Query OK, 218953 rows affected (1.70 sec)
Records: 218953  Duplicates: 0  Warnings: 0

Method 2: Use temporary table

• Create temporary data table tmp_table

mysql> INSERT INTO t_user SELECT * FROM t_user_memory;
Query OK, 218953 rows affected (1.70 sec)
Records: 218953  Duplicates: 0  Warnings: 0

• Use python or bash to generate a 100w recorded data file ( Python will be generated in an instant)

python(推荐): python -c "for i in range(1, 1+1000000): print(i)" > base.txt

• Import data into the temporary table tmp_table

mysql> load data infile '/Users/LJTjintao/temp/base.txt' replace into table tmp_table;
Query OK, 1000000 rows affected (2.55 sec)
Records: 1000000  Deleted: 0  Skipped: 0  Warnings: 0

千万级数据 20秒插入完成

Note : An error may be reported when importing data because mysql does not turn on secure_file_priv by default (This parameter is used to limit the effect of data import and export operations, such as executing LOAD DATA, SELECT... INTO OUTFILE statements and LOAD_FILE() functions. These operations require the user With FILE permission.)

Solution: Add secure_file_priv = /Users/LJTjintao/temp/` in the mysql configuration file (my.ini or my.conf), and then restart mysql solution

• Using the temporary table as the basic data, inserting data into t_user, inserting 100W data takes 10.37s

mysql> INSERT INTO t_user
    ->   SELECT
    ->     id,
    ->     uuid(),
    ->     CONCAT('userNickName', id),
    ->     FLOOR(Rand() * 1000),
    ->     FLOOR(Rand() * 100),
    ->     NOW()
    ->   FROM
    ->     tmp_table;
Query OK, 1000000 rows affected (10.37 sec)
Records: 1000000  Duplicates: 0  Warnings: 0

• Update the creation time field to make the creation time of the inserted data more random

UPDATE t_user SET create_time=date_add(create_time, interval FLOOR(1 + (RAND() * 7)) year);

Query OK, 1000000 rows affected (5.21 sec)
Rows matched: 1000000  Changed: 1000000  Warnings: 0

mysql> UPDATE t_user SET create_time=date_add(create_time, interval FLOOR(1 + (RAND() * 7)) year);


Query OK, 1000000 rows affected (4.77 sec)
Rows matched: 1000000  Changed: 1000000  Warnings: 0

mysql> select * from t_user limit 30;
+----+--------------------------------------+----------------+---------------+-----------+---------------------+
| id | c_user_id                            | c_name         | c_province_id | c_city_id | create_time         |
+----+--------------------------------------+----------------+---------------+-----------+---------------------+
|  1 | bf5e227a-7b84-11e9-9d6e-751d319e85c2 | userNickName1  |            84 |        64 | 2015-11-13 21:13:19 |
|  2 | bf5e26f8-7b84-11e9-9d6e-751d319e85c2 | userNickName2  |           967 |        90 | 2019-11-13 20:19:33 |
|  3 | bf5e2810-7b84-11e9-9d6e-751d319e85c2 | userNickName3  |           623 |        40 | 2014-11-13 20:57:46 |
|  4 | bf5e2888-7b84-11e9-9d6e-751d319e85c2 | userNickName4  |           140 |        49 | 2016-11-13 20:50:11 |
|  5 | bf5e28f6-7b84-11e9-9d6e-751d319e85c2 | userNickName5  |            47 |        75 | 2016-11-13 21:17:38 |
|  6 | bf5e295a-7b84-11e9-9d6e-751d319e85c2 | userNickName6  |           642 |        94 | 2015-11-13 20:57:36 |
|  7 | bf5e29be-7b84-11e9-9d6e-751d319e85c2 | userNickName7  |           780 |         7 | 2015-11-13 20:55:07 |
|  8 | bf5e2a4a-7b84-11e9-9d6e-751d319e85c2 | userNickName8  |            39 |        96 | 2017-11-13 21:42:46 |
|  9 | bf5e2b58-7b84-11e9-9d6e-751d319e85c2 | userNickName9  |           731 |        74 | 2015-11-13 22:48:30 |
| 10 | bf5e2bb2-7b84-11e9-9d6e-751d319e85c2 | userNickName10 |           534 |        43 | 2016-11-13 22:54:10 |
| 11 | bf5e2c16-7b84-11e9-9d6e-751d319e85c2 | userNickName11 |           572 |        55 | 2018-11-13 20:05:19 |
| 12 | bf5e2c70-7b84-11e9-9d6e-751d319e85c2 | userNickName12 |            71 |        68 | 2014-11-13 20:44:04 |
| 13 | bf5e2cca-7b84-11e9-9d6e-751d319e85c2 | userNickName13 |           204 |        97 | 2019-11-13 20:24:23 |
| 14 | bf5e2d2e-7b84-11e9-9d6e-751d319e85c2 | userNickName14 |           249 |        32 | 2019-11-13 22:49:43 |
| 15 | bf5e2d88-7b84-11e9-9d6e-751d319e85c2 | userNickName15 |           900 |        51 | 2019-11-13 20:55:26 |
| 16 | bf5e2dec-7b84-11e9-9d6e-751d319e85c2 | userNickName16 |           854 |        74 | 2018-11-13 22:07:58 |
| 17 | bf5e2e50-7b84-11e9-9d6e-751d319e85c2 | userNickName17 |           136 |        46 | 2013-11-13 21:53:34 |
| 18 | bf5e2eb4-7b84-11e9-9d6e-751d319e85c2 | userNickName18 |           897 |        10 | 2018-11-13 20:03:55 |
| 19 | bf5e2f0e-7b84-11e9-9d6e-751d319e85c2 | userNickName19 |           829 |        83 | 2013-11-13 20:38:54 |
| 20 | bf5e2f68-7b84-11e9-9d6e-751d319e85c2 | userNickName20 |           683 |        91 | 2019-11-13 20:02:42 |
| 21 | bf5e2fcc-7b84-11e9-9d6e-751d319e85c2 | userNickName21 |           511 |        81 | 2013-11-13 21:16:48 |
| 22 | bf5e3026-7b84-11e9-9d6e-751d319e85c2 | userNickName22 |           562 |        35 | 2019-11-13 20:15:52 |
| 23 | bf5e3080-7b84-11e9-9d6e-751d319e85c2 | userNickName23 |            91 |        39 | 2016-11-13 20:28:59 |
| 24 | bf5e30da-7b84-11e9-9d6e-751d319e85c2 | userNickName24 |           677 |        21 | 2016-11-13 21:37:15 |
| 25 | bf5e3134-7b84-11e9-9d6e-751d319e85c2 | userNickName25 |            50 |        60 | 2018-11-13 20:39:20 |
| 26 | bf5e318e-7b84-11e9-9d6e-751d319e85c2 | userNickName26 |           856 |        47 | 2018-11-13 21:24:53 |
| 27 | bf5e31e8-7b84-11e9-9d6e-751d319e85c2 | userNickName27 |           816 |        65 | 2014-11-13 22:06:26 |
| 28 | bf5e324c-7b84-11e9-9d6e-751d319e85c2 | userNickName28 |           806 |         7 | 2019-11-13 20:17:30 |
| 29 | bf5e32a6-7b84-11e9-9d6e-751d319e85c2 | userNickName29 |           973 |        63 | 2014-11-13 21:08:09 |
| 30 | bf5e3300-7b84-11e9-9d6e-751d319e85c2 | userNickName30 |           237 |        29 | 2018-11-13 21:48:17 |
+----+--------------------------------------+----------------+---------------+-----------+---------------------+
30 rows in set (0.01 sec)

More MySQL related For technical articles, please visit the MySQL Tutorial column to learn!

The above is the detailed content of MySQL quickly creates tens of millions of test data. For more information, please follow other related articles on the PHP Chinese website!