mysql json field

Starting from MySQL 5.7.8, MySQL support native JSON data types.

A. Create JSON (length can not be set, may be null, you can not have a default value)

mysql> CREATE TABLE lnmp (
    `id` int(10) unsigned NOT NULL AUTO_INCREMENT,
    `category` JSON,
    `tags` JSON,
    PRIMARY KEY (`id`)
);

mysql> DESC lnmp;
+----------+------------------+------+-----+---------+----------------+
| Field    | Type             | Null | Key | Default | Extra          |
+----------+------------------+------+-----+---------+----------------+
| id       | int(10) unsigned | NO   | PRI | NULL    | auto_increment |
| category | json             | YES  |     | NULL    |                |
| tags     | json             | YES  |     | NULL    |                |
+----------+------------------+------+-----+---------+----------------+
3 rows in set (0.00 sec)

Two insert data

Support dictionaries, lists, and Null, can be directly inserted, a special function can also JSON_OBJECT, JSON_ARRAY json generates data format.

INSERT INTO `lnmp` (category, tags) VALUES ('{"id": 1, "name": "lnmp.cn"}', '[1, 2, 3]');
INSERT INTO `lnmp` (category, tags) VALUES (JSON_OBJECT("id", 2, "name", "php.net"),JSON_ARRAY(1, 3, 5));

See inserted data

mysql> SELECT * FROM lnmp;
+----+------------------------------+-----------+
| id | category                     | tags      |
+----+------------------------------+-----------+
|  1 | {"id": 1, "name": "lnmp.cn"} | [1, 2, 3] |
|  2 | {"id": 2, "name": "php.net"} | [1, 3, 5] |
+----+------------------------------+-----------+
2 rows in set (0.00 sec)

III. Query data

Json query data with the column-> path form, wherein the object type this path represents $ .path , the array type is $ [index]

mysql> SELECT id, category->'$.id', category->'$.name', tags->'$[0]', tags->'$[2]' FROM lnmp;
+----+------------------+--------------------+--------------+--------------+
| id | category->'$.id' | category->'$.name' | tags->'$[0]' | tags->'$[2]' |
+----+------------------+--------------------+--------------+--------------+
|  1 | 1                | "lnmp.cn"          | 1            | 3            |
|  2 | 2                | "php.net"          | 1            | 5            |
+----+------------------+--------------------+--------------+--------------+
2 rows in set (0.00 sec)

Can be seen that the corresponding character string type category -> '. $ Name' also contains a double quotation marks, this is actually not the desired result, can be used to remove double quotes JSON_UNQUOTE function, starting from MySQL 5.7.13 can also the operator - >> this is equivalent to and JSON_UNQUOTE

mysql> SELECT id, category->'$.name', JSON_UNQUOTE(category->'$.name'), category->>'$.name' FROM lnmp;
+----+--------------------+----------------------------------+---------------------+
| id | category->'$.name' | JSON_UNQUOTE(category->'$.name') | category->>'$.name' |
+----+--------------------+----------------------------------+---------------------+
|  1 | "lnmp.cn"          | lnmp.cn                          | lnmp.cn             |
|  2 | "php.net"          | php.net                          | php.net             |
+----+--------------------+----------------------------------+---------------------+
2 rows in set (0.00 sec)

Search as a condition under the following JSON said.

虽然插入的时候可以用字符串的形式插入,但是实际上json字段的存储方式并不等同于字符串,查询的时候直接用字符串相等是不可以的.

可以通过cast方法将字符串转为json

mysql> SELECT * FROM lnmp WHERE category = CAST('{"id": 1, "name": "lnmp.cn"}' as JSON);
+----+------------------------------+-----------+
| id | category                     | tags      |
+----+------------------------------+-----------+
|  1 | {"id": 1, "name": "lnmp.cn"} | [1, 2, 3] |
+----+------------------------------+-----------+
1 row in set (0.00 sec)

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 lnmp WHERE category->'$.name' = 'lnmp.cn';
+----+------------------------------+-----------+
| id | category                     | tags      |
+----+------------------------------+-----------+
|  1 | {"id": 1, "name": "lnmp.cn"} | [1, 2, 3] |
+----+------------------------------+-----------+
1 row in set (0.00 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

除了用 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 lnmp WHERE JSON_CONTAINS(category, '1', '$.id');
+----+------------------------------+-----------+
| id | category                     | tags      |
+----+------------------------------+-----------+
|  1 | {"id": 1, "name": "lnmp.cn"} | [1, 2, 3] |
+----+------------------------------+-----------+
1 row in set (0.01 sec)

JSON_CONTAINS通用于字典查询和列表查询

字典:JSON_CONTAINS(column_name, value, key)

列表：JSON_CONTAINS(column, value)

四.更新数据

如果是整个 json 更新的话，和插入时类似的。

UPDATE lnmp SET tags = '[1, 3, 4]' WHERE id = 1;

更新元素:

 UPDATE lnmp SET category = JSON_INSERT(category, '$.name', 'lnmp', '$.url', 'www.lnmp.cn') WHERE id = 1;
 UPDATE lnmp SET category = JSON_SET(category, '$.host', 'www.lnmp.cn', '$.url', 'http://www.lnmp.cn') WHERE id = 1;
 UPDATE lnmp SET category = JSON_REPLACE(category, '$.name', 'php', '$.url', 'http://www.php.net') WHERE id = 2;
 UPDATE lnmp SET category = JSON_REMOVE(category, '$.url', '$.host') WHERE id = 1;

JSON_INSERT: 只会追加新的值,原有的值不会被覆盖;

JSON_SET: 插入新值并可以覆盖原有的值

JSON_REPLACE: 只替换原有的值,新值忽略

JSON_REMOVE: 删除元素.

五. python应用

from sqlalchemy import JSON
from sqlalchemy import func
self.db().query(TrainCourse).filter(func.json_contains(TrainCourse.teacher_id, "21")).all()
self.db().query(TrainClass).filter(func.json_contains(TrainClass.flex_data, "123", "$.HoldDept")).all()  # 存疑,有时候会报无法理解的错误

与sql原生语句用法类似,可以传三个参数或者两个参数,其中value值必须为字符串.

Guess you like