Mysql JSON对象和JSON数组查询

file_type可以是 varchar，也可以是 json类型

1. 函数说明

JSON_CONTAINS(json_doc, val[, path])：判断是否包含某个json值

JSON_ARRAY([val[, val] ...])：创建json数组

2. JSON对象

{"key": 1, "name": "万飞"}

查询

SELECT * FROM `ak_file_config` where file_type -> '$.name' = "万飞"

3. 字符串JSON数组

3.1 AND关系

["EXE", "白加黑", "DLL"]

查询

SELECT * FROM `ak_file_config` where JSON_CONTAINS(file_type, JSON_ARRAY("白加黑","DLL"))

3.2 OR关系

SELECT * FROM `ak_file_config` where JSON_CONTAINS(file_type,'"DLL"') OR JSON_CONTAINS(file_type,'"EXE"')

4. 对象数组

参考 https://wenku.baidu.com/view/0831b7cc6194dd88d0d233d4b14e852459fb3958?aggId=0831b7cc6194dd88d0d233d4b14e852459fb3958

5. 集成 Mybatis plus

// jsonArray查询
.apply(CollUtil.isNotEmpty(query.getFileType()), StrUtil.format("JSON_CONTAINS(t.file_type, JSON_ARRAY({}))",
                // 设置占位符{0},{1},{2}
                IntStream.range(0, Optional.ofNullable(query.getFileType()).orElse(Collections.emptyList()).size())
                        .mapToObj(i -> "{".concat(String.valueOf(i)).concat("}"))
                        .collect(Collectors.joining(","))),
        Optional.ofNullable(query.getFileType()).orElse(Collections.emptyList()).toArray())

参考 https://blog.csdn.net/qq_31832209/article/details/125374325

6. 模糊查询

SELECT
	* 
FROM
	`ak_file_config` 
WHERE
	JSON_EXTRACT(file_type, '$' ) LIKE '%DL%';

代码里参考

String productOrCompanyName = query.getProductOrCompanyName().replace("\"", "_");

.and(StrUtil.isNotBlank(productOrCompanyName),
                        wq -> wq.apply("JSON_EXTRACT(LOWER(t.label_involve_product), '$') LIKE LOWER(CONCAT('%', {0}, '%'))", productOrCompanyName)
                                .or()
                                .apply("JSON_EXTRACT(LOWER(t.label_involve_company), '$') LIKE LOWER(CONCAT('%', {0}, '%'))", productOrCompanyName))

如果输入字符串带双引号，需要将\"替换成_进行模糊搜索，但是会查询出不带双引号的数据

7. json_table

mysql最低版本8.0.4

7.1 分组计算总数

SELECT
	* 
FROM
	JSON_TABLE ( '["11", "22"]', '$[*]' COLUMNS ( NESTED PATH '$' COLUMNS ( result INT PATH '$' ) ) ) AS t;

json对象数组参考 https://cdn.modb.pro/db/484630

在表中关联字符串数组分组查询

需求：统计各类型的数量

SELECT
	t1.result,
	count( t1.result ) AS count 
FROM
	pe_main_body t
	INNER JOIN JSON_TABLE ( t.overview_product_type, '$[*]' COLUMNS ( NESTED PATH '$' COLUMNS ( result VARCHAR ( 100 ) PATH '$' ) ) ) AS t1 
WHERE
	del_flag = FALSE 
GROUP BY
	t1.result

7.2 对象去重

例如：match_context字段是对象数组，对象里面有两个字段keyWord和describe

需求：查询出所有的不重复的对象

SELECT DISTINCT
	t1.result 
FROM
	ii_sensitive_resource_info t
	INNER JOIN JSON_TABLE ( t.match_context, '$[*]' COLUMNS ( NESTED PATH '$' COLUMNS ( result JSON PATH '$' ) ) ) AS t1 
WHERE
	company_id IN ( 296 )

DISTINCT：去重

改为字段返回

SELECT
	result ->> '$.keyWord' AS keyWord,
	result ->> '$.describe' AS descInfo 
FROM
	(
	SELECT DISTINCT
		t1.result 
	FROM
		ii_sensitive_resource_info t
		INNER JOIN JSON_TABLE ( t.match_context, '$[*]' COLUMNS ( NESTED PATH '$' COLUMNS ( result JSON PATH '$' ) ) ) AS t1 
WHERE
	company_id IN ( 296 )) tmp

->>：会去除双引号

8. JSONArray查询

• JSONArray 符号是 $[0]
• JSONObject 符号是 $

8.1 模糊查询

SELECT * FROM `tf_cloud`.`tf_low_data_testUser` WHERE  `address`->'$[0].name' LIKE "%b%"

8.2 等值匹配

SELECT * FROM `tf_cloud`.`tf_low_data_testUser` WHERE  `address`->'$[0].name' = "bbb"

8.3 时间搜索

SELECT * FROM `tf_cloud`.`tf_low_data_testUser` WHERE CAST(JSON_UNQUOTE(`address`->'$[0].date') AS DATETIME) BETWEEN '2023-08-13' AND '2023-08-17'

8.4 在列表

SELECT * FROM `tf_cloud`.`tf_low_data_testUser` WHERE  `address`->'$[0].name' IN ("bbb","ccc")

8.5 包含

SELECT * FROM `tf_cloud`.`tf_low_data_testUser` WHERE  json_contains(`address`, '["bbb"]', "$[0].nickname")