You can use JSON_TABLE for the same
SELECT get_ddi.*
FROM my_table,
JSON_TABLE(my_column, '$[*]' COLUMNS (
my_column VARCHAR(40) PATH '$.ddi')
) get_ddi;
**Schema (MySQL v8.0)**
DROP TABLE IF EXISTS `my_table`;
CREATE TABLE IF NOT EXISTS `my_table` (
`id` SERIAL,
`my_column` JSON
);
INSERT INTO `my_table` (`id`, `my_column`)
VALUES
(1, '[
{
"ddi": "55",
"routing_id": "7",
"price": {
"mt": 0.0285,
"mo": 0.0285
}
},
{
"ddi": "598",
"routing_id": "10",
"price": {
"mt": 0.06,
"mo": 0.06
}
}
]');
**Query #1**
SELECT get_ddi.*
FROM my_table,
JSON_TABLE(my_column, '$[*]' COLUMNS (
my_column VARCHAR(40) PATH '$.ddi')
) get_ddi;
**Output**
| my_column |
| --------- |
| 55 |
| 598 |
As per @Guilherme Mascarenhas comments below, the solution needed was for MariaDb. As of version 10.2.31, JSON_TABLE function doesn't exist for MariaDB.
A hacky solution could be to use appropriate mariadb sequence table(Depending on the number of rows). JSON_UNQUOTE removes the quote from the extracted value. seq is used as index to get the specified ddi value from the array.
SELECT
JSON_UNQUOTE(JSON_EXTRACT(t.my_column, CONCAT('$[', seq_0_to_100.seq, '].ddi'))) AS getddi
FROM my_table t
JOIN seq_0_to_100
HAVING getddi IS NOT NULL;
**Output**
| my_column |
| --------- |
| 55 |
| 598 |
ddiand another table with one row per price for eachddi. Using JSON is convenient for storing a complex nested structure of data in one step, but your question is an example of where storing data in JSON instead of in normal form makes it much harder to query later.