转载自（我的小碗汤）
MySQL在5.7.8开始对json原生支持，本文将对MySQL中json类型的用法简单说明，希望对你有用。
CREATE TABLE testproject (
`id` int(10) unsigned NOT NULL AUTO_INCREMENT,
`skill` JSON NOT NULL,
`student` JSON NOT NULL,
PRIMARY KEY (`id`)
);
查看表结构：
image
这样JSON的字段就被创建好了
注：JSON类型不能有默认值。
插入JSON
插入 json 格式的字符串，可以是对象的形式，也可以是数组的形式，
INSERT INTO `testproject` (student, skill) VALUES ('{"id": 1, "name": "ggjg"}', '["java", "go", "vue"]');
INSERT INTO `testproject` (student, skill) VALUES ('{"id": 5, "name": "guogege"}', '[]');
插入json时，数据库会对json做校验，不符合json规范就会报错。
image
查询JSON：
查询 json 中的数据用 column->path 的形式，其中对象类型 path 这样表示 .path, 而数组类型则是[index]
查询testproject表student字段中json对象id为1的记录：
SELECT * FROM testproject WHERE student->'$.id'= 1;
image
查询testproject表student字段中json对象id为1或者5的记录：
SELECT * FROM testproject WHERE student->'$.id' in (1,5);
SELECT * FROM testproject WHERE student->'$.id' = 1 or student->'$.id' = 5;
image
也可以用函数json_extract：
image
column->path方法有限制，数据源必须是表字段，否则就报错：
image
以下这样查询，查出来student->'$.name'包含双引号：
SELECT id, student->'$.id', student->'$.name', skill->'$[0]', skill->'$[2]' FROM testproject;
image
这不是我们想要的，可以用 JSON_UNQUOTE 函数将双引号去掉，从 MySQL 5.7.13 起也可以通过这个操作符 ->> 这个和 JSON_UNQUOTE 是等价的。
image
因为 JSON 不同于字符串，所以如果用字符串和 JSON 字段比较，是不会相等的：
mysql> SELECT * FROM testproject WHERE student = '{"id": 1, "name": "ggjg"}';
Empty set (0.00 sec)
此时可以通过 CAST 将字符串转成 JSON 的形式：
mysql> SELECT * FROM testproject WHERE student = CAST('{"id": 1, "name": "ggjg"}' as JSON);
+----+-----------------------+---------------------------+
| id | skill | student |
+----+-----------------------+---------------------------+
| 10 | ["java", "go", "vue"] | {"id": 1, "name": "ggjg"} |
+----+-----------------------+---------------------------+
1 row in set (0.01 sec)
要特别注意的是，JSON 中的元素搜索是严格区分变量类型的，比如说整型和字符串是严格区分的：
mysql> SELECT * FROM testproject WHERE student->'$.id' = '1';
Empty set (0.00 sec)
mysql>
mysql> SELECT * FROM testproject WHERE student->'$.id' = 1;
+----+-----------------------+---------------------------+
| id | skill | student |
+----+-----------------------+---------------------------+
| 10 | ["java", "go", "vue"] | {"id": 1, "name": "ggjg"} |
+----+-----------------------+---------------------------+
1 row in set (0.00 sec)
可以看到搜索字符串 1 和整型 1 的结果是不一样的。
除了用以上 column->path 的形式搜索，还可以用JSON_CONTAINS 函数，但和 column->path 的形式有点相反的是，JSON_CONTAINS 第二个参数是不接受整数的，无论 json 元素是整型还是字符串，否则会出现这个错误：
mysql> SELECT * FROM testproject WHERE JSON_CONTAINS(student, 1, '$.id');
ERROR 3146 (22032): Invalid data type for JSON data in argument 2 to function json_contains; a JSON string or JSON type is required.
mysql>
这里必须要使用字符串：
mysql> SELECT * FROM testproject WHERE JSON_CONTAINS(student, '1', '$.id');
+----+-----------------------+---------------------------+
| id | skill | student |
+----+-----------------------+---------------------------+
| 10 | ["java", "go", "vue"] | {"id": 1, "name": "ggjg"} |
+----+-----------------------+---------------------------+
1 row in set (0.00 sec)
对于数组类型的 JSON 的查询，比如说 skill 中包含有 3 的数据，同样要用 JSON_CONTAINS 函数，同样第二个参数也需要是字符串：
mysql> SELECT * FROM testproject WHERE JSON_CONTAINS(skill, '"go"');
+----+-----------------------+---------------------------+
| id | skill | student |
+----+-----------------------+---------------------------+
| 10 | ["java", "go", "vue"] | {"id": 1, "name": "ggjg"} |
+----+-----------------------+---------------------------+
1 row in set (0.00 sec)
mysql> SELECT * FROM testproject WHERE JSON_CONTAINS(skill, '1');
+----+-----------+------------------------------+
| id | skill | student |
+----+-----------+------------------------------+
| 12 | [1, 2, 3] | {"id": 4, "name": "guogege"} |
+----+-----------+------------------------------+
1 row in set (0.00 sec)
更新数据
MySQL 并不支持 column->path 的形式进行更新操作。
如果是整个 json 更新的话，和插入时类似的：
mysql> select * from testproject where id = 10;
+----+-----------------------+---------------------------+
| id | skill | student |
+----+-----------------------+---------------------------+
| 10 | ["java", "go", "vue"] | {"id": 1, "name": "ggjg"} |
+----+-----------------------+---------------------------+
1 row in set (0.00 sec)
mysql> UPDATE testproject SET skill = '["js", "java"]' WHERE id = 10;
Query OK, 1 row affected (0.01 sec)
Rows matched: 1 Changed: 1 Warnings: 0
mysql> select * from testproject where id = 10;
+----+----------------+---------------------------+
| id | skill | student |
+----+----------------+---------------------------+
| 10 | ["js", "java"] | {"id": 1, "name": "ggjg"} |
+----+----------------+---------------------------+
1 row in set (0.00 sec)
json_array_append和json_array_insert函数使用：
json_array_append是在json后面追加；
json_array_insert是在指定下标插入。
mysql> select * from testproject;
+----+----------------+------------------------------+
| id | skill | student |
+----+----------------+------------------------------+
| 10 | ["js", "java"] | {"id": 1, "name": "ggjg"} |
| 11 | [] | {"id": 5, "name": "guogege"} |
| 12 | [1, 2, 3] | {"id": 4, "name": "guogege"} |
+----+----------------+------------------------------+
3 rows in set (0.00 sec)
mysql> SELECT json_array_append(skill, '$', 'c') from testproject;
+------------------------------------+
| json_array_append(skill, '$', 'c') |
+------------------------------------+
| ["js", "java", "c"] |
| ["c"] |
| [1, 2, 3, "c"] |
+------------------------------------+
3 rows in set (0.00 sec)
mysql> SELECT json_array_insert(skill, '$[1]', 'php') from testproject;
+-----------------------------------------+
| json_array_insert(skill, '$[1]', 'php') |
+-----------------------------------------+
| ["js", "php", "java"] |
| ["php"] |
| [1, "php", 2, 3] |
+-----------------------------------------+
3 rows in set (0.00 sec)
mysql>
json_replace、json_set、json_insert和json_remove函数用法：
json_replace：只替换已经存在的旧值，不存在则忽略；
json_set：替换旧值，并插入不存在的新值；
json_insert：插入新值，但不替换已经存在的旧值；
json_remove() 删除元素。
json_replace：
mysql> select * from testproject;
+----+----------------+--------------------------------+
| id | skill | student |
+----+----------------+--------------------------------+
| 10 | ["js", "java"] | {"id": 1, "name": "smallsoup"} |
| 11 | [] | {"id": 5, "name": "guogege"} |
| 12 | [1, 2, 3] | {"id": 4, "name": "guogege"} |
+----+----------------+--------------------------------+
3 rows in set (0.00 sec)
mysql>
mysql> UPDATE testproject SET student->'$.name' = 'smallsoup' where student->'$.id' = 1;
ERROR 1064 (42000): You have an error in your SQL syntax; check the manual that corresponds to your MySQL server versio
n for the right syntax to use near '->'$.name' = 'smallsoup' where student->'$.id' = 1' at line 1
mysql>
mysql> UPDATE testproject SET student = json_replace(student, '$.name', 'soup') WHERE student->'$.id' = 1;
Query OK, 1 row affected (0.01 sec)
Rows matched: 1 Changed: 1 Warnings: 0
mysql> select * from testproject;
+----+----------------+------------------------------+
| id | skill | student |
+----+----------------+------------------------------+
| 10 | ["js", "java"] | {"id": 1, "name": "soup"} |
| 11 | [] | {"id": 5, "name": "guogege"} |
| 12 | [1, 2, 3] | {"id": 4, "name": "guogege"} |
+----+----------------+------------------------------+
3 rows in set (0.00 sec)
json_set：
mysql> select * from testproject;
+----+----------------+------------------------------+
| id | skill | student |
+----+----------------+------------------------------+
| 10 | ["js", "java"] | {"id": 1, "name": "soup"} |
| 11 | [] | {"id": 5, "name": "guogege"} |
| 12 | [1, 2, 3] | {"id": 4, "name": "guogege"} |
+----+----------------+------------------------------+
3 rows in set (0.00 sec)
mysql> UPDATE testproject SET student = json_set(student, '$.name', 'small', '$.age', 22) WHERE student->'$.id'= 1;
Query OK, 1 row affected (0.01 sec)
Rows matched: 1 Changed: 1 Warnings: 0
mysql> select * from testproject;
+----+----------------+---------------------------------------+
| id | skill | student |
+----+----------------+---------------------------------------+
| 10 | ["js", "java"] | {"id": 1, "age": 22, "name": "small"} |
| 11 | [] | {"id": 5, "name": "guogege"} |
| 12 | [1, 2, 3] | {"id": 4, "name": "guogege"} |
+----+----------------+---------------------------------------+
3 rows in set (0.00 sec)
json_insert：
mysql> select * from testproject;
+----+----------------+---------------------------------------+
| id | skill | student |
+----+----------------+---------------------------------------+
| 10 | ["js", "java"] | {"id": 1, "age": 22, "name": "small"} |
| 11 | [] | {"id": 5, "name": "guogege"} |
| 12 | [1, 2, 3] | {"id": 4, "name": "guogege"} |
+----+----------------+---------------------------------------+
3 rows in set (0.00 sec)
mysql> UPDATE testproject SET student = json_insert(student, '$.name', 'soup', '$.addr', '苏州') WHERE student->'$.id'=
1;
Query OK, 1 row affected (0.00 sec)
Rows matched: 1 Changed: 1 Warnings: 0
mysql> select * from testproject;
+----+----------------+---------------------------------------------------------+
| id | skill | student |
+----+----------------+---------------------------------------------------------+
| 10 | ["js", "java"] | {"id": 1, "age": 22, "addr": "苏州", "name": "small"} |
| 11 | [] | {"id": 5, "name": "guogege"} |
| 12 | [1, 2, 3] | {"id": 4, "name": "guogege"} |
+----+----------------+---------------------------------------------------------+
3 rows in set (0.00 sec)
json_remove() :
mysql> select * from testproject;
+----+----------------+---------------------------------------------------------+
| id | skill | student |
+----+----------------+---------------------------------------------------------+
| 10 | ["js", "java"] | {"id": 1, "age": 22, "addr": "苏州", "name": "small"} |
| 11 | [] | {"id": 5, "name": "guogege"} |
| 12 | [1, 2, 3] | {"id": 4, "name": "guogege"} |
+----+----------------+---------------------------------------------------------+
3 rows in set (0.00 sec)
mysql> UPDATE testproject SET student = json_remove(student, '$.name', '$.age') WHERE student->'$.id' = 1;
Query OK, 1 row affected (0.01 sec)
Rows matched: 1 Changed: 1 Warnings: 0
mysql> select * from testproject;
+----+----------------+------------------------------+
| id | skill | student |
+----+----------------+------------------------------+
| 10 | ["js", "java"] | {"id": 1, "addr": "苏州"} |
| 11 | [] | {"id": 5, "name": "guogege"} |
| 12 | [1, 2, 3] | {"id": 4, "name": "guogege"} |
+----+----------------+------------------------------+
3 rows in set (0.00 sec)
可以看到name和age就被移除了。
以上只列出了部分函数的说明，mysql官方提供的函数列表如下：
image
image
更多用法请查看官方文档：
https://dev.mysql.com/doc/refman/5.7/en/json-function-reference.html