Mysql 展平一个数组

Question

该表包含 json 数据，我想从该 json 中提取 labelKey

表：d_json

data
{"tag":null,"options":[{"labelKey":"key10","value":"yes","selected":true},{"labelKey":"key11","value":"no","selected":false}]}
{"tag":null,"options":[{"labelKey":"key20","value":"yes","selected":true},{"labelKey":"key21","value":"no","selected":false},{"labelKey":"key22","value":"no","selected":false}]}

我使用以下查询来提取“labelKey”

SELECT JSON_EXTRACT(JSON_EXTRACT(j.data,'$.options'),'$[*].labelKey') FROM d_json j AS 结果；

返回如下结果

result
["key10", "key12"]
["key20", "key21", "key22"]

但是我希望结果是平的，每一行包含一个元素而不是数组，例如

result
"key10"
"key11"
"key21"
"key22"
"key23"

不知道如何展平结果数组

Answer 1

在 mysql v8+ 上，您可以使用 JSON_TABLE 函数来做到这一点，如下所示：

SELECT p.*
FROM d_json, 
     JSON_TABLE(data, '$.options[*]' COLUMNS (
                labelKey VARCHAR(40) PATH '$.labelKey')
     ) p;

结果：

labelKey
key10
key11
key20
key21
key22

这是demo fiddle

编辑：

在旧的 MySQL 版本上，试试这个：

SELECT TRIM(SUBSTRING_INDEX(SUBSTRING_INDEX(val2,',',rn),',',-1))
FROM (SELECT 1 rn UNION SELECT 2 UNION SELECT 3 UNION SELECT 4 UNION SELECT 5) AS r
CROSS JOIN
  (SELECT REPLACE(REPLACE(GROUP_CONCAT(val),'[',''),']','') AS val2
     FROM 
       (SELECT JSON_EXTRACT(JSON_EXTRACT(j.data,'$.options'),'$[*].labelKey') AS val
      FROM d_json j) v1 
    ) v2;

Demo fiddle

这个想法是 CROSS JOIN 使用一系列行号，然后使用相同的序列从 GROUP_CONCAT 使用 SUBSTRING_INDEX 的值中提取。在上面的查询示例中，我使用了以下形式的硬编码行序列：

(SELECT 1 rn UNION SELECT 2 UNION SELECT 3 UNION SELECT 4 UNION SELECT 5) AS r

理想情况下，最好的方法是找出所需的序列并动态生成。

更新：

在旧的 MySQL 版本上生成编号序列是一个挑战，尤其是当我们的目标是动态生成时。有一种方法不是动态的，但是可以从一个很长的查询中生成一个很大的编号序列，但是如果你打算长时间使用这个序列，我建议你为它创建一个表：

CREATE TABLE number_seq (
sequences INT);

INSERT INTO number_seq
SELECT @row := @row + 1 AS rn FROM 
(SELECT 0 UNION SELECT 1 UNION SELECT 2 UNION SELECT 3 UNION SELECT 4 UNION 
SELECT 5 UNION SELECT 6 UNION SELECT 7 UNION SELECT 8 UNION SELECT 9) t1 CROSS JOIN
(SELECT 0 UNION SELECT 1 UNION SELECT 2 UNION SELECT 3 UNION SELECT 4 UNION 
SELECT 5 UNION SELECT 6 UNION SELECT 7 UNION SELECT 8 UNION SELECT 9) t2 CROSS JOIN
(SELECT 0 UNION SELECT 1 UNION SELECT 2 UNION SELECT 3 UNION SELECT 4 UNION 
SELECT 5 UNION SELECT 6 UNION SELECT 7 UNION SELECT 8 UNION SELECT 9) t3 CROSS JOIN
(SELECT @row:=0) numbers;

上面的查询将生成 1-1000 范围内的数字并插入到表中。获得该表后，您只需要像这样编写查询：

SELECT TRIM(SUBSTRING_INDEX(SUBSTRING_INDEX(val2,',',sequences),',',-1))
FROM (SELECT sequences FROM 
(SELECT (LENGTH(val2)-LENGTH(REPLACE(val2,',','')))+1 AS valLen FROM
(SELECT REPLACE(REPLACE(GROUP_CONCAT(val),'[',''),']','') AS val2 FROM 
       (SELECT JSON_EXTRACT(JSON_EXTRACT(j.data,'$.options'),'$[*].labelKey') AS val
      FROM d_json j) v1 
    ) v2 ) v3 JOIN number_seq t ON sequences <= valLen) r
CROSS JOIN
  (SELECT REPLACE(REPLACE(GROUP_CONCAT(val),'[',''),']','') AS val2
     FROM 
       (SELECT JSON_EXTRACT(JSON_EXTRACT(j.data,'$.options'),'$[*].labelKey') AS val
      FROM d_json j) v1 
    ) v2;

与上一个查询相比，变化的亮点是硬编码编号序列与查询之间的切换，该查询基本上在最终的JSON_EXTRACT 中获取用逗号分隔的总值，并将其与创建的number_seq 表连接以获得需要的行。这部分在这里：

SELECT sequences FROM 
(SELECT (LENGTH(val2)-LENGTH(REPLACE(val2,',','')))+1 AS valLen FROM
(SELECT REPLACE(REPLACE(GROUP_CONCAT(val),'[',''),']','') AS val2 FROM 
       (SELECT JSON_EXTRACT(JSON_EXTRACT(j.data,'$.options'),'$[*].labelKey') AS val
      FROM d_json j) v1 
    ) v2 ) v3 JOIN number_seq t ON sequences <= valLen

这是一个更新的小提琴供参考https://dbfiddle.uk/?rdbms=mysql_5.7&fiddle=ace8babce8d7bbb97f7e016a754e93a9