1

  1. 我有一个名为postscolumns的表id, title, tags。

     CREATE TABLE IF NOT EXISTS `posts` (
   `id` int(6) unsigned NOT NULL,
   `title` varchar(100) NOT NULL,
   `tags` json NOT NULL,
   PRIMARY KEY (`id`)
 ) DEFAULT CHARSET=utf8;

  2. [{"tag": "android", "time": 122}, {"tag": "apple", "time": 140}]像在tags字段中一样存储数据。

      INSERT INTO `posts` (`id`, `title`, `tags`) VALUES
       ('1', 'First Post', '[{"tag": "andoroid", "time": 123}, {"tag": "mobiles", "time": 432} ]'),
       ('2', 'Second Post', '[{"tag": "apple", "time": 125}]'),
       ('3', 'Third Post', '[{"tag": "android", "time": 124}]'),
       ('4', 'Fourth Post', '[{"tag": "mobiles", "time": 472}, {"tag": "android", "time": 129}]'),
       ('5', 'Fifth Post', '[{"tag": "android", "time": 122}, {"tag": "apple", "time": 140}]'),
       ('6', 'Sixth Post', '[{"tag": "mobiles", "time": 121}, {"tag": "apple", "time": 120}]'),
       ('7', 'Seventh  Post', '[{"tag": "apple", "time": 120}, {"tag": "mobiles", "time": 130}]'),
       ('8', 'Eigth  Post', '[{"tag": "android", "time": 126}]'),
       ('9', 'Nineth  Post', '[{"tag": "mobiles", "time":132}]');

  3. 根据标签值过滤表中的数据，例如。tag == "android". 为了实现这一点，我正在使用 mysql 查询

    SELECT id, title, tags FROM posts where JSON_CONTAINS(
                 tags , '{"tag": "android"}'
                 ) ;

它工作正常。DB小提琴：https ://www.db-fiddle.com/f/5Hw1FyL3Qv2RLtCw4tZbyh/1

此外，我需要根据timetags 中的值订购结果。tag == 'android' and order by time.

提前感谢您的帮助。

4

1 回答 1

3

在 MySQL 8 之前的版本中解决这个问题的唯一方法是使用存储函数time从每组标签中找到最小值。像这样的东西：

DELIMITER //
CREATE FUNCTION min_time(tags JSON)
RETURNS INT
BEGIN
  DECLARE i INT DEFAULT 0;
  DECLARE mint INT DEFAULT 9999;
  DECLARE thist INT;
  SET thist = JSON_EXTRACT(tags, CONCAT('$[', i, '].time'));
  WHILE thist IS NOT NULL DO
    IF thist < mint THEN
      SET mint = thist;
    END IF;
    SET i = i + 1;
    SET thist = JSON_EXTRACT(tags, CONCAT('$[', i, '].time'));
  END WHILE;
  RETURN mint;
END //

然后你可以使用这样的查询：

SELECT id, title, tags, min_time(tags) AS min_time
FROM posts 
WHERE JSON_CONTAINS(tags , '{"tag": "android"}')
ORDER BY min_time

输出：

id  title       tags                                                                tags->'$[*].time'    min_time
4   Fourth Post [{"tag": "mobiles", "time": 472}, {"tag": "android", "time": 121}]  [472, 121]          121
5   Fifth Post  [{"tag": "android", "time": 122}, {"tag": "apple", "time": 140}]    [122, 140]          122
3   Third Post  [{"tag": "android", "time": 124}]                                   [124]               124
8   Eigth Post  [{"tag": "android", "time": 126}]                                   [126]               126

dbfiddle 上的演示

这可能是过度设计的，因为它按帖子上任何标签的最短时间排序。如果您只想android按与标签（您正在搜索的那个）关联的时间排序，您可以使用这个简化的查询：

SELECT id, title, tags,
       JSON_EXTRACT(tags, CONCAT(SUBSTRING_INDEX(JSON_UNQUOTE(JSON_SEARCH(tags, 'one', 'android')), '.', 1), '.time')) AS tag_time
FROM posts 
WHERE JSON_CONTAINS(tags , '{"tag": "android"}')
ORDER BY tag_time

输出：

id  title       tags                                                                tag_time
3   Third Post  [{"tag": "android", "time": 75}]                                    75
4   Fourth Post [{"tag": "mobiles", "time": 472}, {"tag": "android", "time": 121}]  121
5   Fifth Post  [{"tag": "android", "time": 122}, {"tag": "apple", "time": 140}]    122
8   Eigth Post  [{"tag": "android", "time": 126}]                                   126

dbfiddle 上的演示

于 2019-04-10T23:58:07.130 回答