【MySQL】字符集不一樣致使的問題

一、兩表字符集不一樣致使沒法使用索引

開發反映有個請求sql很慢，須要3s以上。sql以下：

SELECT
  tc.customer_name,
  cb.service_record_id~~~~,
  tc.customer_code,
  tc.customer_desc,
  tc.customer_account_id,
  tc.phone
FROM
  tur_customer tc
INNER JOIN tur_customer_bind cb ON tc.customer_account_id = cb.customer_account_id
WHERE
  cb.employee_account_id = '181'
AND tc.is_delete = 0
ORDER BY
  cb.service_record_id DESC
LIMIT 0,10

表結構：

CREATE TABLE `tur_customer` (
  `id` int(11) unsigned NOT NULL AUTO_INCREMENT COMMENT '主鍵',
  `customer_code` varchar(64) NOT NULL COMMENT '客戶編碼',
  `customer_name` varchar(255) NOT NULL COMMENT '客戶姓名',
  `customer_name_hash` varchar(128) DEFAULT NULL COMMENT '客戶姓名Hash',
  `customer_account_id` varchar(32) NOT NULL COMMENT '賬號系統對應的惟一標識',
  `real_name_status` int(11) DEFAULT '10' COMMENT '實名制認證狀態: 10 未認證, 20 已實名, 30 已認證',
  `sex` int(11) DEFAULT '1' COMMENT '性別;1男；0女',
  `phone` varchar(128) NOT NULL COMMENT '聯繫電話，可能時手機號，也多是座機號碼',
  PRIMARY KEY (`id`),
  UNIQUE KEY `uniq_customercode` (`customer_code`) USING BTREE,
  UNIQUE KEY `uniq_customeraccountid` (`customer_account_id`) USING BTREE,
  KEY `idx_customernamehash` (`customer_name_hash`) USING BTREE
) ENGINE=InnoDB AUTO_INCREMENT=267031 DEFAULT CHARSET=utf8 COMMENT='客戶基本信息表';
 
CREATE TABLE `tur_customer_bind` (
  `id` int(11) unsigned NOT NULL AUTO_INCREMENT COMMENT 'id',
  `customer_account_id` varchar(32) NOT NULL COMMENT '客戶帳號',
  `employee_account_id` varchar(32) NOT NULL COMMENT '顧問帳號',
  `employee_account_name` varchar(32) NOT NULL COMMENT '顧問名稱',
  `service_record_id` int(11) NOT NULL COMMENT '最新服務記錄ID',
  `bind_time` datetime NOT NULL COMMENT '綁定時間',
  `store_code` varchar(32) DEFAULT NULL COMMENT '店面編碼',
  PRIMARY KEY (`id`) USING BTREE,
  UNIQUE KEY `uniq_account` (`customer_account_id`) USING BTREE COMMENT '客戶帳號惟一索引',
  KEY `idx_employee_account_id` (`employee_account_id`) USING BTREE COMMENT '顧問帳號索引'
) ENGINE=InnoDB AUTO_INCREMENT=14583 DEFAULT CHARSET=utf8mb4 ROW_FORMAT=COMPACT COMMENT='當前客戶關係表';

分析表的執行計劃以下：

+----+-------------+-------+------+--------------------------------------+-------------------------+---------+-------+-------+--------------------------------------------------------+
| id | select_type | table | type | possible_keys                        | key                     | key_len | ref   | rows  | Extra                                                  |
+----+-------------+-------+------+--------------------------------------+-------------------------+---------+-------+-------+--------------------------------------------------------+
|  1 | SIMPLE      | cb    | ref  | uniq_account,idx_employee_account_id | idx_employee_account_id | 130     | const |     1 | Using index condition; Using temporary; Using filesort |
|  1 | SIMPLE      | tc    | ALL  | NULL                                 | NULL                    | NULL    | NULL  | 58697 | Using where; Using join buffer (Block Nested Loop)     |
+----+-------------+-------+------+--------------------------------------+-------------------------+---------+-------+-------+--------------------------------------------------------+

兩表關聯字段tur_customer.customer_account_id和tur_customer_bind.customer_account_id都有索引，且類型都爲varchar，可是索引沒有用到，查看字符集發現一個表是utf8一個是utf8mb4。

兩表的字符集不一樣，會形成沒法使用索引的狀況發生。 查看庫的字符集爲utf8，排序規則爲utf8_general_ci.

將表tur_customer修改成utf8mb4格式後，該sql的執行時間僅爲0.1s，再次查看執行計劃以下：

+----+-------------+-------+--------+--------------------------------------+-------------------------+---------+-----------------------------+------+----------------------------------------------------+
| id | select_type | table | type   | possible_keys                        | key                     | key_len | ref                         | rows | Extra                                              |
+----+-------------+-------+--------+--------------------------------------+-------------------------+---------+-----------------------------+------+----------------------------------------------------+
|  1 | SIMPLE      | cb    | ref    | uniq_account,idx_employee_account_id | idx_employee_account_id | 130     | const                       |    1 | Using index condition; Using where; Using filesort |
|  1 | SIMPLE      | tc    | eq_ref | uniq_customeraccountid               | uniq_customeraccountid  | 130     | test.cb.customer_account_id |    1 | Using where                                        |
+----+-------------+-------+--------+--------------------------------------+-------------------------+---------+-----------------------------+------+----------------------------------------------------+

二、兩表排序規則不一樣進行join連表報錯

報錯信息以下：

SQL state [HY000]; error code [1267]; Illegal mix of collations (utf8mb4_general_ci,IMPLICIT) and (utf8mb4_unicode_ci,IMPLICIT) for operation '='; nested exception is java.sql.SQLException: Illegal mix of collations (utf8mb4_general_ci,IMPLICIT) and (utf8mb4_unicode_ci,IMPLICIT) for operation '='"

查了下兩表的 DDL，一個是utf8mb4_unicode_ci，一個是utf8mb4_general_ci，因此致使了這個問題

解決方法： 將這兩個表統一改成 utf8mb4_unicode_ci 或者 utf8mb4_general_ci。

若是不修改表結構，能夠在查詢的時候指定等號一邊的字符集編碼:

• 格式：WHERE 列名 COLLATE utf8mb4_unicode_ci = 嵌套語句
• 或者：WHERE 列名 = 嵌套語句 COLLATE utf8mb4_unicode_ci 達到讓兩邊的字符集相等！