MySQL5.7之JSON使用
MySQL 支持原生的 JSON 数据类型，列不能设置默认值
从MySQL 5.7.8 开始，MySQL支持原生的JSON格式，即有独立的json类型，用于存放 json格式的数据。
JSON 格式的数据并不是以string格式存储于数据库而是以内部的binary 格式，以便于快速的定位到json 格式中值
在插入和更新操作时MySQL会对JSON 类型做校验，已检查数据是否符合json格式，如果不符合则报错。
同时5.7.8 版本提供了四种JSON相关的函数，从而不用遍历全部数据
a 创建: JSON_ARRAY(), JSON_MERGE(), JSON_OBJECT()
b 修改: JSON_APPEND(), JSON_ARRAY_APPEND(), JSON_ARRAY_INSERT(), JSON_INSERT(), JSON_QUOTE(), JSON_REMOVE(), JSON_REPLACE(),
JSON_SET(), and JSON_UNQUOTE()
c 查询: JSON_CONTAINS(), JSON_CONTAINS_PATH(), JSON_EXTRACT(), JSON_KEYS(),JSON_SEARCH().
d 属性: JSON_DEPTH(), JSON_LENGTH(), JSON_TYPE() JSON_VALID().
CREATE TABLE t1(id INT PRIMARY KEY AUTO_INCREMENT,jdoc json)
1、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 t1(jdoc)VALUES('{"key1":"value1","key2":"value2"}')
INSERT INTO t1(jdoc)VALUE('[1,2,3]')
2、查看插入的JSON数据类型:json_type()
SELECT json_type(jdoc) FROM t1
mysql> select json_type('"abc"') ;
+--------------------+
| json_type('"abc"') |
+--------------------+
| STRING |
+--------------------+
mysql> select * from t1;
+----+--------------------------------------+
| id | jdoc |
+----+--------------------------------------+
| 1 | {"key1": "value1", "key2": "value2"} |
| 2 | [1, 2, 3] |
+----+--------------------------------------+
2 rows in set (0.00 sec)
mysql> select json_type(jdoc) from t1;
+-----------------+
| json_type(jdoc) |
+-----------------+
| OBJECT |
| ARRAY |
+-----------------+
mysql> select json_type('["a","b",1]');
+--------------------------+
| json_type('["a","b",1]') |
+--------------------------+
| ARRAY |
+--------------------------+
1 row in set (0.00 sec)
3、用专门的函数将字符串变成JSON数组格式存储，json_object()函数根据键值对返回JSON对象
json_array
mysql> select json_array("a",1,now());
+----------------------------------------+
| json_array("a",1,now()) |
+----------------------------------------+
| ["a", 1, "2017-11-02 01:47:38.000000"] |
+----------------------------------------+
json_object
mysql> select json_object("key1",1,"key2","abc");
+------------------------------------+
| json_object("key1",1,"key2","abc") |
+------------------------------------+
| {"key1": 1, "key2": "abc"} |
+------------------------------------+
1 row in set (0.00 sec)
json_merge():将两个或两个以上的JSON文档合并成一个列表对象
mysql> select json_merge('["a",1,2,3]','{"key1":"value"}');
+----------------------------------------------+
| json_merge('["a",1,2,3]','{"key1":"value"}') |
+----------------------------------------------+
| ["a", 1, 2, 3, {"key1": "value"}] |
+----------------------------------------------+
若是多个列表的JSON格式会被合并成一个列表存储
mysql> SELECT JSON_MERGE('[1, 2]', '["a", "b"]', '[true, false]');
+-----------------------------------------------------+
| JSON_MERGE('[1, 2]', '["a", "b"]', '[true, false]') |
+-----------------------------------------------------+
| [1, 2, "a", "b", true, false] |
+-----------------------------------------------------+
1 row in set (0.00 sec)
若多个对象有相同的键，值会合并成一个数组。
4、将JSON值自定义一个变量
mysql> set @a=json_object("key1","sharesoe");
Query OK, 0 rows affected (0.00 sec)
mysql> select @a;
+----------------------+
| @a |
+----------------------+
| {"key1": "sharesoe"} |
+----------------------+
1 row in set (0.00 sec)
5、查询和修改JSON值
通过提取JSON中键获取值
mysql> select json_extract(jdoc,'$.key1'),json_extract(jdoc,'$.key2')from t1;
+-----------------------------+-----------------------------+
| json_extract(jdoc,'$.key1') | json_extract(jdoc,'$.key2') |
+-----------------------------+-----------------------------+
| "value1" | "value2" |
+-----------------------------+-----------------------------+
1 row in set (0.00 sec)
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 $[index]
mysql> select id,jdoc->'$.key1',jdoc->'$.key2' from t1;
+----+----------------+----------------+
| id | jdoc->'$.key1' | jdoc->'$.key2' |
+----+----------------+----------------+
| 1 | "value1" | "value2" |
+----+----------------+----------------+
可以用 JSON_UNQUOTE 函数将双引号去掉，从 MySQL 5.7.13 起也可以通过这个操作符 ->> 这个和 JSON_UNQUOTE 是等价的
mysql> select id,jdoc->>'$.key1',jdoc->>'$.key2' from t1;
+----+-----------------+-----------------+
| id | jdoc->>'$.key1' | jdoc->>'$.key2' |
+----+-----------------+-----------------+
| 1 | value1 | value2 |
+----+-----------------+-----------------+
1 row in set (0.00 sec)
mysql> select id,json_unquote(jdoc->'$.key1'),jdoc->'$.key2' from t1;
+----+------------------------------+----------------+
| id | json_unquote(jdoc->'$.key1') | jdoc->'$.key2' |
+----+------------------------------+----------------+
| 1 | value1 | "value2" |
+----+------------------------------+----------------+
JSON 作为条件进行搜索。因为 JSON 不同于字符串，所以如果用字符串和 JSON 字段比较，是不会相等的
mysql> select * from t1 where jdoc='{"key1":"value1","key2":"value2"}';
Empty set (0.00 sec)
可以通过 CAST 将字符串转成 JSON 的形式
mysql> select * from t1 where jdoc=cast('{"key1":"value1","key2":"value2"}' as json);
+----+--------------------------------------+
| id | jdoc |
+----+--------------------------------------+
| 1 | {"key1": "value1", "key2": "value2"} |
+----+--------------------------------------+
通过 JSON 中的元素进行查询, 对象型的查询同样可以通过 column->path
mysql> select * from t1 where jdoc->'$.key1'="value1";
+----+--------------------------------------+
| id | jdoc |
+----+--------------------------------------+
| 1 | {"key1": "value1", "key2": "value2"} |
+----+--------------------------------------+
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
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，若第2个位字符也是有问题
select * from t1 where JSON_CONTAINS(jdoc,"value1","$.key1");
更新 JSON
如果是整个 json 更新的话，和插入时类似的。但如果要更新 JSON 下的元素，MySQL 并不支持 column->path 的形式
JSON_INSERT() 插入新值，但不会覆盖已经存在的值
mysql> update t1 set jdoc=json_insert(jdoc,'$.key1','abcd','$.name','www.sharesoe.com') where id=1;
Query OK, 1 row affected (0.11 sec)
mysql> select * from t1;
+----+------------------------------------------------------------------+
| id | jdoc |
+----+------------------------------------------------------------------+
| 1 | {"key1": "value1", "key2": "value2", "name": "www.sharesoe.com"} |
+----+------------------------------------------------------------------+
1 row in set (0.00 sec)
JSON_SET() 插入新值，并覆盖已经存在的值
mysql> update t1 set jdoc=json_set(jdoc,'$.key1','abcd','$.host','www.sharesoe.com') where id=1;
Query OK, 1 row affected (0.39 sec)
mysql> select * from t1;
+----+--------------------------------------------------------------------------------------------+
| id | jdoc |
+----+--------------------------------------------------------------------------------------------+
| 1 | {"host": "www.sharesoe.com", "key1": "abcd", "key2": "value2", "name": "www.sharesoe.com"} |
+----+--------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)
JSON_REPLACE() 只替换存在的值
mysql> update t1 set jdoc=json_replace(jdoc,'$.key1','abcd1','$.host1','www.sharesoe.com') where id=1;
Query OK, 1 row affected (0.14 sec)
mysql> select * from t1;
+----+---------------------------------------------------------------------------------------------+
| id | jdoc |
+----+---------------------------------------------------------------------------------------------+
| 1 | {"host": "www.sharesoe.com", "key1": "abcd1", "key2": "value2", "name": "www.sharesoe.com"} |
+----+---------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)
JSON_REMOVE() 删除 JSON 元素
mysql> update t1 set jdoc=json_remove(jdoc,'$.key1','$.host') where id=1;
Query OK, 1 row affected (0.28 sec)
mysql> select * from t1;
+----+------------------------------------------------+
| id | jdoc |
+----+------------------------------------------------+
| 1 | {"key2": "value2", "name": "www.sharesoe.com"} |
+----+------------------------------------------------+
1 row in set (0.00 sec)
对于使用MySQL varchar 存储json的业务是一个福音。在设计表的时候，避免大字段存储。
一来是减少不必要的查询带来的IO，带宽，内存方面的影响 二来是 避免因为表大小太大导致的ddl 时间成本增加系统风险。
