1

我们可以有很多请求将 json 存储到数据库。有时我们需要将 JSON 键集转换为表 - 行。如何在Mysql表中提取完整的json？

即具有以下json和要求是将每个集合存储为表中的一行。

'{
"log": [
{
"datetime": "2017-10-25 07:19:55",
"Activity": "Activity 1"
},
{
"datetime": "2017-10-25 07:20:05",
"Activity": "Activity 2"
},
{
"datetime": "2017-10-25 07:20:31",
"Activity": "Activity 3"
},
{
"datetime": "2017-10-25 07:20:31",
"Activity": "Activity 4"
},
{
"datetime": "2017-10-25 07:21:03",
"Activity": "Activity 5"
},
{
"datetime": "2017-10-25 13:56:42",
"Activity": "Activity 6"
},
{
"datetime": "2017-10-25 13:56:53",
"Activity": "Activity 7"
},
{
"datetime": "2017-10-25 13:57:03",
"Activity": "Activity 8"
},
{
"datetime": "2017-10-25 13:57:04",
"Activity": "Activity 9"
}, 
{ "datetime": "2017-10-25 13:57:53",
"Activity": "Activity 10"
},
{
"datetime": "2017-10-25 13:57:59",
"Activity": "Activity 11"
},
{
"datetime": "2017-10-25 13:58:01",
"Activity": "Activity 12"
},
{
"datetime": "2017-10-25 13:58:01",
"Activity": "Activity 13"
}
]
} '
4

2 回答 2

1

MySQL 8.0 有一个 JSON_TABLE 函数：

INSERT INTO NewTable 
  SELECT * FROM JSON_TABLE('...your JSON ...', '$.log[*]' columns(
    rowid for ordinality, 
    dt datetime path '$.datetime', 
    activity varchar(20) path '$.Activity')
  ) as t
于 2018-12-26T06:40:48.813 回答
0

JSON 可以存储在具有不同数据类型的数据库中。即 JSON、VACRHAR、TEXT...等。对于导出报告或过滤 JSON 元素值，我们应该有不同的解决方案来完成它。

创建一个表以将 JSON 存储在 MySQL 表中。

CREATE TABLE test_Json(id INT, activity_log JSON);

将样本数据插入表中。JSON 字段数据以 JSON 数据类型插入，但也可以是 VARCHAR 或 Text 类型。

INSERT INTO test_Json(id, activity_log)
VALUES(1,
'{
"log": [
{
"datetime": "2017-10-25 07:19:55",
"Activity": "Activity 1"
},
{
"datetime": "2017-10-25 07:20:05",
"Activity": "Activity 2"
},
{
"datetime": "2017-10-25 07:20:31",
"Activity": "Activity 3"
},
{
"datetime": "2017-10-25 07:20:31",
"Activity": "Activity 4"
},
{
"datetime": "2017-10-25 07:21:03",
"Activity": "Activity 5"
},
{
"datetime": "2017-10-25 13:56:42",
"Activity": "Activity 6"
},
{
"datetime": "2017-10-25 13:56:53",
"Activity": "Activity 7"
},
{
"datetime": "2017-10-25 13:57:03",
"Activity": "Activity 8"
},
{
"datetime": "2017-10-25 13:57:04",
"Activity": "Activity 9"
}, { "datetime": "2017-10-25 13:57:53",
"Activity": "Activity 10"
},
{
"datetime": "2017-10-25 13:57:59",
"Activity": "Activity 11"
},
{
"datetime": "2017-10-25 13:58:01",
"Activity": "Activity 12"
},
{
"datetime": "2017-10-25 13:58:01",
"Activity": "Activity 13"
}
]
} ');

INSERT INTO test_Json(id, activity_log)
VALUES(2,'
{
"log": [
{ "datetime": "2017-10-29 21:56:27",
"Activity": "Activity-11"
},
{
"datetime": "2017-10-29 21:56:27",
"Activity": "Activity-12"
},
{
"datetime": "2017-10-29 21:56:27",
"Activity": "Activity-13"
},
{
"datetime": "2017-10-29 21:56:27",
"Activity": "Activity-14"
},
{
"datetime": "2017-10-29 21:56:35",
"Activity": "Activity-15"
},
{
"datetime": "2017-10-29 21:56:49",
"Activity": "Activity-16"
},
{
"datetime": "2017-10-29 21:56:49",
"Activity": "Activity-17"
},
{
"datetime": "2017-10-29 21:56:49",
"Activity": "Activity-18"
}
]
}');

查询以将 JSON 导出为具有 JSON 长度值的 JSON 数组

SELECT TJ.id,
JSON_LENGTH(JSON_EXTRACT(TJ.activity_log, '$**.datetime')) as Json_Length,
JSON_EXTRACT(TJ.activity_log, '$**.datetime') as json_array_message,
JSON_EXTRACT(TJ.activity_log, '$**.Activity') as json_array_datetime
FROM test_Json TJ
WHERE JSON_VALID(TJ.activity_log) = 1;

JSON_VALID()将验证 JSON 列数据，无论 JSON 数据是否正确。如果不是，那么它将从查询中过滤掉。

递归 CTE（公用表表达式）以行格式获取 n 级元素值

WITH RECURSIVE CTE_json (id, array_datetime, array_message,seq) AS
(
SELECT id, 
JSON_EXTRACT(JSON_EXTRACT(activity_log, '$**.datetime'), CONCAT('$[', 0, ']')), 
JSON_EXTRACT(JSON_EXTRACT(activity_log, '$**.Activity'), CONCAT('$[', 0, ']')), 
0 as seq
FROM test_Json
UNION ALL
SELECT TJ.id, 
JSON_EXTRACT(JSON_EXTRACT(TJ.activity_log, '$**.datetime'), CONCAT('$[', rc.seq +1, ']')) as array_datetime, 
JSON_EXTRACT(JSON_EXTRACT(TJ.activity_log, '$**.Activity'), CONCAT('$[', rc.seq +1, ']')) as array_message, 
rc.seq +1 AS seq
FROM CTE_json rc
INNER JOIN test_Json TJ ON rc.id = TJ.id
WHERE rc.seq+1 < JSON_LENGTH(JSON_EXTRACT(TJ.activity_log, '$**.datetime'))
)

SELECT *
FROM CTE_json cj
ORDER BY id, seq;

如果当前 MySQL 版本不支持递归 CTE，那么我们可以使用其他 LOOP 替代方案。

于 2018-12-26T05:32:39.307 回答