json - 将 JSON 数组上的 MariaDB 表连接到文本字段

Question

表crons有一个名为的 JSON 字段surveys，其中包含一个 UUID 的 JSON 数组，例如：

[
  "65f42c322ea442099612cad251be1dac",
  "43788fe207b143dc9eac4da38c036af6"
]

表1（计划）：

CREATE TABLE "crons" (
  "id" int NOT NULL AUTO_INCREMENT,
  "depts" longtext CHARACTER SET utf8mb4 COLLATE utf8mb4_bin,
  "surveys" longtext CHARACTER SET utf8mb4 COLLATE utf8mb4_bin,
  PRIMARY KEY ("id"),
  CONSTRAINT "CONSTRAINT_1" CHECK (((`depts` is null) or json_valid(`depts`))),
  CONSTRAINT "CONSTRAINT_2" CHECK (((`surveys` is null) or json_valid(`surveys`))),
  CONSTRAINT "crons_chk_1" CHECK (json_valid(`depts`)),
  CONSTRAINT "crons_chk_2" CHECK (json_valid(`surveys`))
)

表2（调查）：

表surveys包含survey_uuid作为 varchar 字符串的字段。

CREATE TABLE "surveys" (
  "id" int NOT NULL AUTO_INCREMENT,
  "survey_uuid" varchar(32) NOT NULL,
  PRIMARY KEY ("id"),
  UNIQUE KEY "surveys_survey_uuid_uindex" ("survey_uuid"),
  CONSTRAINT "survey_json" CHECK (((`survey_json` is null) or json_valid(`survey_json`)))
)

我需要返回surveys.survey_uuid(varchar) 包含在crons.surveys(JSON array)中的所有行

试图：

SELECT *
FROM crons c
         JOIN
     surveys s
    ON (JSON_SEARCH(c.surveys, 'one', s.survey_uuid) IS NOT NULL)
WHERE start_date < NOW()
  AND end_date > NOW()
  AND JSON_SEARCH(c.depts, 'one', 'sales') IS NOT NULL

注意：MariaDB 将 JSON 存储为长文本，但所有 JSON 函数都有效。我已经验证这JSON_SEARCH(depts, 'one', 'sales') IS NOT NULL是搜索 JSON 字段数组的有效方法。

在 MariaDB 中你不能做的一件事是，CAST(something AS JSON)但你可以CAST(jsonfield AS char)

Answer 1

如果要搜索 中surveys至少有一个匹配行的所有行crons，则应使用以下 SQL：

SELECT s.id, s.survey_uuid, c.id, c.depts, c.surveys
  FROM surveys s JOIN crons c
  ON (JSON_SEARCH(c.surveys, 'all', s.survey_uuid) IS NOT NULL)
  ORDER BY s.id;

如果要包含surveys其中没有行的行crons，请使用LEFT JOIN.

---

使用以下数据：

MariaDB [test]> select * from crons;
+----+--------------+--------------------------------------------------------------+
| id | depts        | surveys                                                      |
+----+--------------+--------------------------------------------------------------+
|  1 | {"d1": "d1"} | ["B8606CA2F4000004"]                                         |
|  2 | {"d2": "d2"} | ["B8606CA2F4000005"]                                         |
|  3 | {"d3": "d3"} | ["B8606CA2F4000004", "B8606CA2F4000006"]                     |
|  4 | {"d4": "d4"} | ["B8606CA2F4000004", "B8606CA2F4000005", "B8606CA2F4000007"] |
|  5 | {"d5": "d5"} | [""]                                                         |
|  6 | {"d6": "d6"} | [""]                                                         |
+----+--------------+--------------------------------------------------------------+
6 rows in set (0.000 sec)

MariaDB [test]> select * from surveys;
+----+------------------+
| id | survey_uuid      |
+----+------------------+
|  5 | B8606CA2F4000004 |
|  6 | B8606CA2F4000005 |
|  7 | B8606CA2F4000006 |
|  8 | B8606CA2F4000007 |
|  9 | B8606CA2F4000008 |
| 10 | B8606CA2F4000009 |
| 11 | B8606CA2F400000A |
+----+------------------+
7 rows in set (0.000 sec)

查询的结果是：

MariaDB [test]> SELECT s.id, s.survey_uuid, c.id, c.depts, c.surveys FROM surveys s JOIN crons c ON (JSON_SEARCH(c.surveys, 'all', s.survey_uuid) IS NOT NULL) ORDER BY s.id;
+----+------------------+----+--------------+--------------------------------------------------------------+
| id | survey_uuid      | id | depts        | surveys                                                      |
+----+------------------+----+--------------+--------------------------------------------------------------+
|  5 | B8606CA2F4000004 |  1 | {"d1": "d1"} | ["B8606CA2F4000004"]                                         |
|  5 | B8606CA2F4000004 |  3 | {"d3": "d3"} | ["B8606CA2F4000004", "B8606CA2F4000006"]                     |
|  5 | B8606CA2F4000004 |  4 | {"d4": "d4"} | ["B8606CA2F4000004", "B8606CA2F4000005", "B8606CA2F4000007"] |
|  6 | B8606CA2F4000005 |  2 | {"d2": "d2"} | ["B8606CA2F4000005"]                                         |
|  6 | B8606CA2F4000005 |  4 | {"d4": "d4"} | ["B8606CA2F4000004", "B8606CA2F4000005", "B8606CA2F4000007"] |
|  7 | B8606CA2F4000006 |  3 | {"d3": "d3"} | ["B8606CA2F4000004", "B8606CA2F4000006"]                     |
|  8 | B8606CA2F4000007 |  4 | {"d4": "d4"} | ["B8606CA2F4000004", "B8606CA2F4000005", "B8606CA2F4000007"] |
+----+------------------+----+--------------+--------------------------------------------------------------+
7 rows in set (0.000 sec)