程序员需了解的SQL之JSON函数查询与修改（七）

文章目录

前言

MySQL从5.7.8以后引入了JSON数据类型，对于JSON文档的操作除了简单的读和写之外还有很多处理JSON的函数。

JSON函数可以从增删改查这些功能点来深入学习Mysql提供的JSON函数.JSON支持包括NUMBER、STRING、BOOLEAN、NULL、ARRAY、OBJECT共6种。

1.创建JSON函数

创建JSON数组

mysql> select json_array(1,'json',true,null,now());
+-------------------------------------------------------+
| json_array(1,'json',true,null,now())                  |
+-------------------------------------------------------+
| [1, "json", true, null, "2021-02-21 22:19:57.000000"] |
+-------------------------------------------------------+
1 row in set (0.01 sec)

创建JSON对象

mysql> select json_object('name','张三','age',24);
+---------------------------------------+
| json_object('name','张三','age',24)   |
+---------------------------------------+
| {
   
   "age": 24, "name": "张三"}           |
+---------------------------------------+
1 row in set (0.00 sec)

2. 查询JSON函数

2.1 JSON_CONAINS(source,target,path)

此函数可以判断target是否包含在source中，其中path参数可选，如果有参数为NULL或者path不存在则返回NULL，存在返回1否则返回0。
下面以简单例子演示此函数使用：

## 创建一个JSON数组对象
mysql> select json_array(1,2,'abc',true,null);
+---------------------------------+
| json_array(1,2,'abc',true,null) |
+---------------------------------+
| [1, 2, "abc", true, null]       |
+---------------------------------+
1 row in set (0.00 sec)

然后判断元素abc、true、null、520是否存在上述JSON对象中

判断元素abc是否包含JSON对象中

mysql> select json_contains('[1, 2, "abc", true, null]','"abc"');
+----------------------------------------------------+
| json_contains('[1, 2, "abc", true, null]','"abc"') |
+----------------------------------------------------+
|                                                  1 |
+----------------------------------------------------+
1 row in set (0.00 sec)

判断元素true是否包含JSON对象中

mysql> select json_contains('[1, 2, "abc", true, null]','true');
+---------------------------------------------------+
| json_contains('[1, 2, "abc", true, null]','true') |
+---------------------------------------------------+
|                                                 1 |
+---------------------------------------------------+
1 row in set (0.00 sec)

判断元素null是否包含在JSON对象中

mysql> select json_contains('[1, 2, "abc", true, null]','null');
+---------------------------------------------------+
| json_contains('[1, 2, "abc", true, null]','null') |
+---------------------------------------------------+
|                                                 1 |
+---------------------------------------------------+
1 row in set (0.00 sec)

判断元素520是否包含在此函数中

mysql> select json_contains('[1, 2, "abc", true, null]','520');
+--------------------------------------------------+
| json_contains('[1, 2, "abc", true, null]','520') |
+--------------------------------------------------+
|                                                0 |
+--------------------------------------------------+
1 row in set (0.00 sec)

判断数组[1,2,‘abc’] 是否存在JSON

mysql> select json_contains('[1, 2, "abc", true, null]','[1,2,"abc"]');
+----------------------------------------------------------+
| json_contains('[1, 2, "abc", true, null]','[1,2,"abc"]') |
+----------------------------------------------------------+
|                                                        1 |
+----------------------------------------------------------+
1 row in set (0.00 sec)

上面几个案例中我们使用JSON_CONTAINS 没有使用到path 参数，下面我们将演示带path参数的函数使用。
通常path以$.key,如果key是一个对象obj，则path应该为$.obj.key。

创建JSON对象并存储在Mysql变量中

mysql> select json_object('name','张三','age',23);
+---------------------------------------+
| json_object('name','张三','age',23)   |
+---------------------------------------+
| {
   
   "age": 23, "name": "张三"}           |
+---------------------------------------+
1 row in set (0.00 sec)
## 创建name1变量
mysql> set @name1='{"age": 23, "name": "张三"}';
Query OK, 0 rows affected (0.00 sec)

判断路径为$.name 是否存在值张三，包含返回1不包含返回0。

mysql> select json_contains(@name1,'"张三"','$.name');
+-------------------------------------------+
| json_contains(@name1,'"张三"','$.name')   |
+-------------------------------------------+
|                                         1 |
+-------------------------------------------+
1 row in set (0.00 sec)

2.2 JSON_CONAINS_PATH(source,one or all,path…)

此函数查询指定的多个path参数是否存在JSON文档中，其中第二个参数只能取ONE或者ALL，分别表示匹配一个就可以以及匹配所有。

## 匹配JSON对象包含指定路径name或者xx
mysql> select json_contains_path(@name1,'one','$.name','$.xx');
+--------------------------------------------------+
| json_contains_path(@name1,'one','$.name','$.xx') |
+--------------------------------------------------+
|                                                1 |
+--------------------------------------------------+
1 row in set (0.00 sec)
## 匹配JSON对象包含指定路径name和xx
mysql> select json_contains_path(@name1,'all','$.name','$.age');
+---------------------------------------------------+
| json_contains_path(@name1,'all','$.name','$.age') |
+---------------------------------------------------+
|                                                 1 |
+---------------------------------------------------+
1 row in set (0.00 sec)

2.3 JSON_EXTRACT(source,path…)

在JSON数组中可以path可以$[index] 表示数组中第index个元素，下面提取第index个元素如下：

## 创建JSON数组
mysql> select json_array('1','2',json_array('3','4'));
+-----------------------------------------+
| json_array('1','2',json_array('3','4')) |
+-----------------------------------------+
| ["1", "2", ["3", "4"]]                  |
+-----------------------------------------+
1 row in set (0.00 sec)

抽取第一个和第三个值

mysql> select json_extract('["1", "2", ["3", "4"]]','$[0]','$[2]');
+------------------------------------------------------+
| json_extract('["1", "2", ["3", "4"]]','$[0]','$[2]') |
+------------------------------------------------------+
| ["1", ["3", "4"]]                                    |
+------------------------------------------------------+
1 row in set (0.00 sec)

MySQL5.7.9之后有一个更简单的标号-> 用来简化JSON_EXRACT,其语法如下：

列名 -> path

## 创建包含JSON类型的表如下
mysql> create table json_demo(content json);
Query OK, 0 rows affected (0.01 sec)
## 插入数据
mysql> insert into json_demo values(@name1);
Query OK, 1 row affected (0.01 sec)
## 查询数据
mysql> select * from json_demo;
+-------------------------------+
| content                       |
+-------------------------------+
| {
   
   "age": 23, "name": "张三"}   |
+-------------------------------+
1 row in set (0.00 sec)

请看如下SQL语句：

mysql> select content,json_extract(content,'$.name') from json_demo where json_extract(content,'$.name')='张三';
+-------------------------------+--------------------------------+
| content                       | json_extract(content,'$.name') |
+-------------------------------+--------------------------------+
| {
   
   "age": 23, "name": "张三"}   | "张三"                         |
+-------------------------------+--------------------------------+
1 row in set (0.00 sec)

等价于如下SQL：

mysql> select content,content -> '$.name' from json_demo where content -> '$.name'='张三';
+-------------------------------+---------------------+
| content                       | content -> '$.name' |
+-------------------------------+---------------------+
| {
   
   "age": 23, "name": "张三"}   | "张三"              |
+-------------------------------+---------------------+
1 row in set (0.00 sec)

JSON中查询为字符串类型都会带双引号，如果去掉双引号可以使用JSON_UNQUOTE 函数如下所示：

mysql> select content,json_unquote(content -> '$.name') from json_demo where content -> '$.name'='张三';
+-------------------------------+-----------------------------------+
| content                       | json_unquote(content -> '$.name') |
+-------------------------------+-----------------------------------+
| {
   
   "age": 23, "name": "张三"}   | 张三                              |
+-------------------------------+-----------------------------------+
1 row in set (0.01 sec)

2.4JSON_KEYS(source,path)

此函数返回在指定路径下所有的键，下面简单演示其使用：

mysql> select json_keys(@name1);
+-------------------+
| json_keys(@name1) |
+-------------------+
| ["age", "name"]   |
+-------------------+
1 row in set (0.00 sec)
mysql>

创建嵌套JSON对象如下所示：

mysql> select json_object('name','张三','age','23','account',json_object('accountId','123456','accountBank','招商银行'));
+------------------------------------------------------------------------------------------------------------------+
| json_object('name','张三','age','23','account',json_object('accountId','123456','accountBank','招商银行'))       |
+------------------------------------------------------------------------------------------------------------------+
| {
   
   "age": "23", "name": "张三", "account": {
   
   "accountId": "123456", "accountBank": "招商银行"}}                     |
+------------------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

查询嵌套函数key

mysql> select json_keys(@name2);
+----------------------------+
| json_keys(@name2)          |
+----------------------------+
| ["age", "name", "account"] |
+----------------------------+
1 row in set (0.00 sec)

查询指定路径下嵌套的key

mysql> select json_keys(@name2,'$.account');
+-------------------------------+
| json_keys(@name2,'$.account') |
+-------------------------------+
| ["accountId", "accountBank"]  |
+-------------------------------+
1 row in set (0.00 sec)

2.5 JSON_SEARCH(source,one or all ,search_str)

此函数表示在source中匹配一个或者所有复合search_str的元素key，其中search_str 可使用类似like中模糊匹配如下：

## 查询JSON至少一个内容是23的key
mysql> select json_search(@name2,'one','23');
+--------------------------------+
| json_search(@name2,'one','23') |
+--------------------------------+
| "$.age"                        |
+--------------------------------+
1 row in set (0.00 sec)
## 查询JSON至少一个内容是以张开头的key
mysql> select json_search(@name2,'one','张%');
+----------------------------------+
| json_search(@name2,'one','张%')  |
+----------------------------------+
| "$.name"                         |
+----------------------------------+
1 row in set (0.00 sec)
## 查询内容包含2的所有key对象
mysql> select json_search(@name2,'all','%2%');
+----------------------------------+
| json_search(@name2,'all','%2%')  |
+----------------------------------+
| ["$.age", "$.account.accountId"] |
+----------------------------------+
1 row in set (0.00 sec)
## 创建新的JSON对象并赋值
mysql> select json_array('1','2',true,null,'abc');
+-------------------------------------+
| json_array('1','2',true,null,'abc') |
+-------------------------------------+
| ["1", "2", true, null, "abc"]       |
+-------------------------------------+
1 row in set (0.00 sec)
mysql> set @v1='["1", "2", true, null, "abc"]';
Query OK, 0 rows affected (0.00 sec)
## 查询JSON数组
mysql> select json_search(@v1,'one','1');
+----------------------------+
| json_search(@v1,'one','1') |
+----------------------------+
| "$[0]"                     |
+----------------------------+
1 row in set (0.00 sec)

3. 修改JSON函数

3.1 JSON_ARRAY_APPEND(source,path,val)

此函数在指定path的JSON 数组尾部加val。如果指定path是一个JSON对象则将其封装城一个新的JSON Array。

mysql> select json_array_append(@v1,'$[0]','haha');
+-----------------------------------------+
| json_array_append(@v1,'$[0]','haha')    |
+-----------------------------------------+
| [["1", "haha"], "2", true, null, "abc"] |
+-----------------------------------------+
1 row in set (0.00 sec)

mysql> select json_array_append('[["1", "haha"], "2", true, null, "abc"]','$[0]','xixi');
+----------------------------------------------------------------------------+
| json_array_append('[["1", "haha"], "2", true, null, "abc"]','$[0]','xixi') |
+----------------------------------------------------------------------------+
| [["1", "haha", "xixi"], "2", true, null, "abc"]                            |
+----------------------------------------------------------------------------+
1 row in set (0.00 sec)
## 设置值
mysql> set @v2='[["1", "haha", "xixi"], "2", true, null, "abc"]';
Query OK, 0 rows affected (0.00 sec)
## 在第一个元素数组中第二个位置上追加 ai
mysql> select json_array_append(@v2,'$[0][1]','ai');
+---------------------------------------------------------+
| json_array_append(@v2,'$[0][1]','ai')                   |
+---------------------------------------------------------+
| [["1", ["haha", "ai"], "xixi"], "2", true, null, "abc"] |
+---------------------------------------------------------+
1 row in set (0.00 sec)

3.2 JSON_ARRAY_INSERT(source,path,val)

此函数可以在指定path下插入元素val，原先位置的元素整体右移。如果插入的目标元素非JSON数组则不会插入val元素，如果插入的元素超出JSON对象的长度则在尾部进行追加。

mysql> select json_array_insert(@name2,'$[0]','1');
+----------------------------------------------------------------------------------------------------+
| json_array_insert(@name2,'$[0]','1')                                                               |
+----------------------------------------------------------------------------------------------------+
| {
   
   "age": "23", "name": "张三", "account": {
   
   "accountId": "123456", "accountBank": "招商银行"}}       |
+----------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

上面的JSON对象插入没有效果，这是因为JSON对象是一个JSON_OBJECT类型的元素会自动忽略插入的元素‘1’。
接下来我们使用JSON_ARRAY插入对象看一看效果。

mysql> select json_array_insert('["1", "2", ["3", "4"]]','$[0]','x');
+--------------------------------------------------------+
| json_array_insert('["1", "2", ["3", "4"]]','$[0]','x') |
+--------------------------------------------------------+
| ["x", "1", "2", ["3", "4"]]                            |
+--------------------------------------------------------+
1 row in set (0.00 sec)

3.3 JSON_REPLACE(source,path,val)

此函数可以替换指定路径上的数据，如果path路径不存在则忽略。

mysql> select json_replace(@v1,'$[0]','y');
+------------------------------+
| json_replace(@v1,'$[0]','y') |
+------------------------------+
| ["y", "1", "2", ["3", "4"]]  |
+------------------------------+
1 row in set (0.00 sec)


mysql> select json_replace(@name2,'$.age','24');
+----------------------------------------------------------------------------------------------------+
| json_replace(@name2,'$.age','24')                                                                  |
+----------------------------------------------------------------------------------------------------+
| {
   
   "age": "24", "name": "张三", "account": {
   
   "accountId": "123456", "accountBank": "招商银行"}}       |
+----------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

3.4 JSON_SET(source,path,val)

此函数可以设置指定路径的数据，与JSON_REPLACE功能类似，唯一区别就是当指定路径不存在，会在尾部进行添加

mysql> select json_replace(@name2,'$.age11','24');
+----------------------------------------------------------------------------------------------------+
| json_replace(@name2,'$.age11','24')                                                                |
+----------------------------------------------------------------------------------------------------+
| {
   
   "age": "23", "name": "张三", "account": {
   
   "accountId": "123456", "accountBank": "招商银行"}}       |
+----------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

我们发现此函数并不存在，我们在使用JSON_SET函数在看次效果如下：

mysql> select json_set(@name2,'$.age11','24');
+-------------------------------------------------------------------------------------------------------------------+
| json_set(@name2,'$.age11','24')                                                                                   |
+-------------------------------------------------------------------------------------------------------------------+
| {
   
   "age": "23", "name": "张三", "age11": "24", "account": {
   
   "accountId": "123456", "accountBank": "招商银行"}}       |
+-------------------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

3.5 JSON_MERGE_PRESERVE(source1,source2)

此函数是可以合并两个JSON函数，合成规则如下：

如果两者都是JSON数组，则合成一个JSON数组对象。
如果都是JSON对象则合成一个JSON对象。
如果是不同类型，则将非JSON数组对象转换成JSON数组在按照第一个规则进行合并

合并数组

mysql> select json_merge_preserve(@v1,json_array('a','b','c'));
+--------------------------------------------------+
| json_merge_preserve(@v1,json_array('a','b','c')) |
+--------------------------------------------------+
| ["x", "1", "2", ["3", "4"], "a", "b", "c"]       |
+--------------------------------------------------+
1 row in set (0.00 sec)

合并对象

mysql> select json_merge_preserve(@name2,json_object('company','神码'));
+-------------------------------------------------------------------------------------------------------------------------+
| json_merge_preserve(@name2,json_object('company','神码'))                                                               |
+-------------------------------------------------------------------------------------------------------------------------+
| {
   
   "age": "23", "name": "张三", "account": {
   
   "accountId": "123456", "accountBank": "招商银行"}, "company": "神码"}         |
+-------------------------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

合并多类型

mysql> select json_merge_preserve(@v1,@name2);
+---------------------------------------------------------------------------------------------------------------------------------+
| json_merge_preserve(@v1,@name2)                                                                                                 |
+---------------------------------------------------------------------------------------------------------------------------------+
| ["x", "1", "2", ["3", "4"], {
   
   "age": "23", "name": "张三", "account": {
   
   "accountId": "123456", "accountBank": "招商银行"}}]       |
+---------------------------------------------------------------------------------------------------------------------------------+
1 row in set (0.00 sec)

3.5 JSON_REMOVE(source,path…)

此函数可以移除指定路径的数据，如果某个路径不存在则略过此路径。

mysql> select json_remove(@v1,'$[0]','$[1]');
+--------------------------------+
| json_remove(@v1,'$[0]','$[1]') |
+--------------------------------+
| ["1", ["3", "4"]]              |
+--------------------------------+
1 row in set (0.00 sec)

这里需要注意的是删除是串行操作的，即先删除$[0]后的JSON文档上基础继续删除$[1]上的元素。