Tag Archives: Json

MySQL5.7で実装されたJSONデータ型とJSON関数のレビューになります。
ご利用になる場合は、参考にして頂ければと思います。

ネイティブJSONデータ型 (バイナリ形式)
Insert時のJSON構文バリデーション機能
組み込みJSON関数 (保存、検索、更新、操作)
ドキュメントにインデックス設定し高速アクセス
SQLとの統合を容易にする、新しいインライン構文
utf8mb4の文字セットとutf8mb4_binの照合 「🐬」
サイズはmax_allowed_packetの値で制限 (Default:4MB)

MySQL5.7からは、リレーショナル、スキーマレスを同じ技術スタックで利用可能になっています。

13.16.1 JSON Function Reference
https://dev.mysql.com/doc/refman/5.7/en/json-function-reference.html

参考：　Modifying JSON Values in MySQL 5.7
https://planet.mysql.com/entry/?id=5994648

JSONで表現する全てのデータ型をサポート

数値, 文字列, bool(true,false)
オブジェクト {“キー”: “値”}, 配列 [123456, “String”, …]
null
日付(date), 時刻, 日付(datetime), タイムスタンプ, その他

[CONFIRM]> show create table T_JSON_SUPPORT\G
*************************** 1. row ***************************
       Table: T_JSON_SUPPORT
Create Table: CREATE TABLE `T_JSON_SUPPORT` (
`id` int(10) NOT NULL AUTO_INCREMENT,
`body` json DEFAULT NULL,
`type` varchar(20) GENERATED ALWAYS AS (json_type(`body`)) VIRTUAL,
  PRIMARY KEY (`id`)
) ENGINE=InnoDB AUTO_INCREMENT=8 DEFAULT CHARSET=utf8mb4


[CONFIRM]> select * from T_JSON_SUPPORT;
+----+---------------------------------------------+----------+
| id | body                                        | type     |
+----+---------------------------------------------+----------+
|  1 | 123456789                                   | INTEGER  |
|  2 | NULL                                        | NULL     |
|  3 | true                                        | BOOLEAN  |
|  4 | "abcde"                                     | STRING   |
|  5 | {"id": 5, "name": "オブジェクト"}            | OBJECT   |
|  6 | [-122.42200352825247, 37.80848009696725, 0] | ARRAY    |
|  7 | "2016-02-29"                                | DATE     |
|  8 | "2016-02-29 00:00:00.000000"                | DATETIME |
+----+---------------------------------------------+----------+

データ型と照合順
4byteなので絵文字を格納可能、またutf8mb4_binなので大文字と小文字を区別します。

 [NEW57]> SET @j = JSON_OBJECT('key', 'value');
1 row in set (0.00 sec)

 [NEW57]> SELECT @j;
+------------------+
| @j               |
+------------------+
| {"key": "value"} |
+------------------+
1 row in set (0.00 sec)

 [NEW57]> SELECT CHARSET(@j), COLLATION(@j);
+-------------+---------------+
| CHARSET(@j) | COLLATION(@j) |
+-------------+---------------+
| utf8mb4     | utf8mb4_bin   |
+-------------+---------------+
1 row in set (0.00 sec)

JSONドキュメントと生成列を利用して列を作成し、対象列にINDEXを付けて高速な検索を行う事が可能です
生成列はファンクションインデックスとして利用可能

こちらは、生成列を利用したテーブルの作成例になります。
string1とstring2にデータをINSERTする事で、string1_w_string、string2_w_string、compareのデータは自動生成されます。

[CONFIRM]> show create table T_Character_COLLATE_utf8mb4_bin\G
*************************** 1. row ***************************
       Table: T_Character_COLLATE_utf8mb4_bin
Create Table: CREATE TABLE `T_Character_COLLATE_utf8mb4_bin` (
  `pid` int(10) unsigned NOT NULL AUTO_INCREMENT,
  `string1` char(1) CHARACTER SET utf8mb4 COLLATE utf8mb4_bin DEFAULT NULL,
  `string2` char(1) CHARACTER SET utf8mb4 COLLATE utf8mb4_bin DEFAULT NULL,
  `string1_w_string` char(8) GENERATED ALWAYS AS (hex(weight_string(`string1`))) VIRTUAL,
  `string2_w_string` char(8) GENERATED ALWAYS AS (hex(weight_string(`string2`))) VIRTUAL,
  `compare` char(1) GENERATED ALWAYS AS ((`string1` = `string2`)) VIRTUAL,
  PRIMARY KEY (`pid`)
) ENGINE=InnoDB AUTO_INCREMENT=1 DEFAULT CHARSET=utf8mb4

 [CONFIRM]> insert into T_Character_COLLATE_utf8mb4_bin(string1,string2) values('A','a');
Query OK, 1 row affected (0.00 sec)

 [CONFIRM]> insert into T_Character_COLLATE_utf8mb4_bin(string1,string2) values('あ','ぁ');
Query OK, 1 row affected (0.00 sec)

 [CONFIRM]> insert into T_Character_COLLATE_utf8mb4_bin(string1,string2) values('A','A');
Query OK, 1 row affected (0.00 sec)

 [CONFIRM]> select * from T_Character_COLLATE_utf8mb4_bin;
+-----+---------+---------+------------------+------------------+---------+
| pid | string1 | string2 | string1_w_string | string2_w_string | compare |
+-----+---------+---------+------------------+------------------+---------+
|   1 | A       | a       | 000041           | 000061           | 0       |
|   2 | あ      | ぁ      | 003042           | 003041           | 0       |
|   3 | A       | A       | 000041           | 000041           | 1       |
+-----+---------+---------+------------------+------------------+---------+

JSONデータと生成列を利用すると、特定のJSON識別子からデータを抜き出して列を作成し、その列にインデックスを付ける事が可能
例えば、JSONデータをINSERTしてJSONドキュメントにUSER_IDやPRODUCT_IDがあれば、その項目だけを抜き出して列にしてINDEXを付与。
検索する場合は、それらの値を利用して検索するとJSONドキュメントも高速検索が可能になります。

[NEW57]> CREATE TABLE `T_JSON` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `feature` json NOT NULL,
  `feature_type` varchar(30) GENERATED ALWAYS AS (json_unquote(feature->"$.type")) VIRTUAL,
  `feature_street` varchar(30) GENERATED ALWAYS AS (json_extract(`feature`,‘$.properties.STREET’)) VIRTUAL,
  PRIMARY KEY (`id`),
  KEY `idx_feature_street` (`feature_street`)
) ENGINE=InnoDB AUTO_INCREMENT=1 DEFAULT CHARSET=utf8mb4;


[NEW57]> alter table features add index idx_feature_type(`feature_type`);
Query OK, 0 rows affected (6.14 sec)
Records: 0  Duplicates: 0  Warnings: 0

[NEW57]> explain select distinct(feature_type) from features;
+----+-------------+----------+------------+-------+------------------+------------------+---------+------+--------+----------+-------------+
| id | select_type | table    | partitions | type  | possible_keys    | key              | key_len | ref  | rows   | filtered | Extra       |
+----+-------------+----------+------------+-------+------------------+------------------+---------+------+--------+----------+-------------+
|  1 | SIMPLE      | features | NULL       | index | idx_feature_type | idx_feature_type | 123     | NULL | 199013 |   100.00 | Using index |
+----+-------------+----------+------------+-------+------------------+------------------+---------+------+--------+----------+-------------+

【JSON関数の使い方】

(1)JSONドキュメントの情報取得

[NEW57]> SELECT JSON_VALID(body) from T_JSON_DOC where id = 5;
+------------------+
| JSON_VALID(body) |
+------------------+
|                1 |
+------------------+

[NEW57]> SELECT JSON_TYPE(body) from T_JSON_DOC where id = 5;
+-----------------+
| JSON_TYPE(body) |
+-----------------+
| OBJECT          |
+-----------------+

[NEW57]> SELECT JSON_KEYS(body) from T_JSON_DOC where id = 5;
+---------------------------------------+
| JSON_KEYS(body)                       |
+---------------------------------------+
| ["id", "name", "price", "Conditions"] |
+---------------------------------------+


[NEW57]> SELECT JSON_SEARCH(feature,'one','MARKET') AS extract_path FROM features WHERE id = 121254;
+-----------------------+
| extract_path          |
+-----------------------+
| "$.properties.STREET" |
+-----------------------+

(2)通常のリレーショナルテーブルから,JSONドキュメントを作成する
Object{}とArray[]関数を利用

/*** リレーショナルテーブル ***/
[NEW57]> SELECT NAME,CountryCode from world.City where CountryCode ='JPN' limit 1;
+-------+-------------+
| NAME  | CountryCode |
+-------+-------------+
| Tokyo | JPN         |
+-------+-------------+

/*** リレーショナルテーブルからJSONデータを作成 ({})***/
[NEW57]> SELECT JSON_OBJECT('CITY',NAME,'Country',CountryCode) from world.City where CountryCode ='JPN' limit 1;
+------------------------------------------------+
| JSON_OBJECT('CITY',NAME,'Country',CountryCode) |
+------------------------------------------------+
| {"CITY": "Tokyo", "Country": "JPN"}            |
+------------------------------------------------+

/*** VIEWにしておくと呼び出しがより楽になります ***/
[NEW57]> CREATE VIEW v_City_json AS
    -> SELECT JSON_OBJECT('ID', ID, 'name', Name, 'CountryCode', CountryCode, 'District', District,'Population',Population) as doc FROM City where CountryCode = 'JPN';
Query OK, 0 rows affected (0.01 sec)

[NEW57]> select * from v_City_json limit 1\G
*************************** 1. row ***************************
doc: {"ID": 1532, "name": "Tokyo", "District": "Tokyo-to", "Population": 7980230, "CountryCode": "JPN"}
1 row in set (0.01 sec)


/*** リレーショナルテーブル ***/
[NEW57]>SELECT NAME,CountryCode from world.City where CountryCode ='JPN' limit 1;
+-------+-------------+
| NAME  | CountryCode |
+-------+-------------+
| Tokyo | JPN         |
+-------+-------------+


/*** リレーショナルテーブルからJSONデータを作成 ([])***/
[NEW57]> SELECT JSON_ARRAY(NAME,CountryCode)  from world.City where CountryCode ='JPN' limit 1;
+------------------------------+
| JSON_ARRAY(NAME,CountryCode) |
+------------------------------+
| ["Tokyo", "JPN"]             |
+------------------------------+


/*** リレーショナルテーブル***/
[NEW57]> select * from T_JSON_PLACE;
+----+-------------------------------------+----------+-----------+
| id | place                               | latitude | longitude |
+----+-------------------------------------+----------+-----------+
|  1 | 東京都港区北青山2-5-8                | 35.67125 | 139.71864 |
|  2 | 大阪府大阪市北区堂2-4－27           　| 34.69584 | 135.49291 |
+----+-------------------------------------+----------+-----------+


/*** リレーショナルテーブルからJSONデータを作成 ({},[])***/
[NEW57]> select JSON_OBJECT('ID',id,'住所',place,'緯度経度', JSON_ARRAY(latitude,longitude)) from T_JSON_PLACE;
+---------------------------------------------------------------------------------------------------+
| JSON_OBJECT('ID',id,'住所',place,'緯度経度', JSON_ARRAY(latitude,longitude))                      |
+---------------------------------------------------------------------------------------------------+
| {"ID": 1, "住所": "東京都港区北青山2-5-8", "緯度経度": [35.67125, 139.71864]}                     |
| {"ID": 2, "住所": "大阪府大阪市北区堂2-4－27", "緯度経度": [34.69584, 135.49291]}                 |
+---------------------------------------------------------------------------------------------------+

メモ：　ST_AsGeoJSONやST_GeomFromGeoJSON等のSpatial GeoJSON Functionsで空間を表すJSONも利用する事が可能

JSONドキュメントの更新処理

/**** JSON_REMOVE (SELECTでの確認とUPDATEによる更新)****/

[NEW57]> select * from T_JSON_DOC;
+----+---------------------------------------------------------------------------------------------------+------------+
| id | body                                                                                              | price_only |
+----+---------------------------------------------------------------------------------------------------+------------+
|  1 | {"id": 1, "name": "自転車", "price": 10000, "Conditions": ["NEW", 2015, "Excellent"]}             |   10000.00 |
|  2 | {"id": 2, "name": "テレビ", "price": 30000, "Conditions": ["USED", 2013, "故障"]}                 |   30000.00 |
|  3 | {"id": 3, "name": "冷蔵庫", "price": 17131, "Conditions": ["NEW", 2015]}                          |   17131.00 |
|  4 | {"id": 4, "name": "オートバイ", "price": 500000, "Conditions": ["NEW", 2015]}                     |  500000.00 |
|  5 | {"id": 5, "name": "自転車", "price": 25000, "Quantity": "[1,10]", "Conditions": ["NEW", 2015]}    |   25000.00 |
+----+---------------------------------------------------------------------------------------------------+------------+

[NEW57]> select JSON_REMOVE(body,"$.Quantity") from T_JSON_DOC where id = 5;
+-----------------------------------------------------------------------------+
| JSON_REMOVE(body,"$.Quantity")                                              |
+-----------------------------------------------------------------------------+
| {"id": 5, "name": "自転車", "price": 25000, "Conditions": ["NEW", 2015]}    |
+-----------------------------------------------------------------------------+

[NEW57]> update T_JSON_DOC set body = JSON_REMOVE(body,"$.Quantity") where id = 5;



/**** JSON_SET (SELECTでの確認とUPDATEによる更新)****/

[NEW57]> select * from T_JSON_DOC;
+----+------------------------------------------------------------------------------+------------+
| id | body                                                                         | price_only |
+----+------------------------------------------------------------------------------+------------+
|  1 | {"id": 1, "name": "自転車", "price": 10000, "Conditions": ["NEW", 2015]}     |   10000.00 |
|  2 | {"id": 2, "name": "テレビ", "price": 30000, "Conditions": ["USED",2013]}     |   30000.00 |
|  3 | {"id": 3, "name": "冷蔵庫", "price": 11154, "Conditions": ["NEW", 2015]}     |   11154.00 |
|  4 | {"id": 4, "name": "冷蔵庫", "price": 50000, "Conditions": ["NEW", 2015]}     |   50000.00 |
|  5 | {"id": 5, "name": "自転車", "price": 25000, "Conditions": ["NEW", 2015]}     |   25000.00 |
+----+------------------------------------------------------------------------------+------------+

[NEW57]> SELECT JSON_SET(body,'$.name',"オートバイ", '$.price',500000) from T_JSON_DOC where id = 4;
+------------------------------------------------------------------------------------+
| JSON_SET(body,'$.name',"オートバイ", '$.price',500000)                             |
+------------------------------------------------------------------------------------+
| {"id": 4, "name": "オートバイ", "price": 500000, "Conditions": ["NEW", 2015]}      |
+------------------------------------------------------------------------------------+

[NEW57]> update T_JSON_DOC set body = JSON_SET(body,'$.name',"オートバイ", '$.price',500000) where id = 4;


/**** JSON_INSERT (SELECTでの確認とUPDATEによる更新)****/

[NEW57]> select * from T_JSON_DOC;
+----+------------------------------------------------------------------------------+------------+
| id | body                                                                         | price_only |
+----+------------------------------------------------------------------------------+------------+
|  1 | {"id": 1, "name": "自転車", "price": 10000, "Conditions": ["NEW", 2015]}     |   10000.00 |
|  2 | {"id": 2, "name": "テレビ", "price": 30000, "Conditions": ["USED",2013]}     |   30000.00 |
|  3 | {"id": 3, "name": "冷蔵庫", "price": 11154, "Conditions": ["NEW", 2015]}     |   11154.00 |
|  4 | {"id": 4, "name": "冷蔵庫", "price": 50000, "Conditions": ["NEW", 2015]}     |   50000.00 |
|  5 | {"id": 5, "name": "自転車", "price": 25000, "Conditions": ["NEW", 2015]}     |   25000.00 |
+----+------------------------------------------------------------------------------+------------+

[NEW57]> select JSON_INSERT(body,'$.Quantity','[1,10]') from NEW57.T_JSON_DOC where id = 5;
+---------------------------------------------------------------------------------------------------+
| JSON_INSERT(body,'$.Quantity','[1,10]')                                                           |
+---------------------------------------------------------------------------------------------------+
| {"id": 5, "name": "自転車", "price": 25000, "Quantity": "[1,10]", "Conditions": ["NEW", 2015]}    |
+---------------------------------------------------------------------------------------------------+

[NEW57]> update NEW57.T_JSON_DOC set body = JSON_INSERT(body,'$.Quantity','[1,10]') where id = 5;


/**** JSON_REPLACE (SELECTでの確認とUPDATEによる更新)****/

[NEW57]> select * from T_JSON_DOC;
+----+------------------------------------------------------------------------------+------------+
| id | body                                                                         | price_only |
+----+------------------------------------------------------------------------------+------------+
|  1 | {"id": 1, "name": "自転車", "price": 10000, "Conditions": ["NEW", 2015]}     |   10000.00 |
|  2 | {"id": 2, "name": "テレビ", "price": 30000, "Conditions": ["USED",2013]}     |   30000.00 |
|  3 | {"id": 3, "name": "冷蔵庫", "price": 11154, "Conditions": ["NEW", 2015]}     |   11154.00 |
|  4 | {"id": 4, "name": "冷蔵庫", "price": 50000, "Conditions": ["NEW", 2015]}     |   50000.00 |
|  5 | {"id": 5, "name": "自転車", "price": 25000, "Conditions": ["NEW", 2015]}     |   25000.00 |
+----+------------------------------------------------------------------------------+------------+

[NEW57]> select JSON_REPLACE(body,"$.price",FLOOR(10000 + (RAND() * 9000))) from T_JSON_DOC where id = 3;
+-----------------------------------------------------------------------------+
| JSON_REPLACE(body,"$.price",FLOOR(10000 + (RAND() * 9000)))                 |
+-----------------------------------------------------------------------------+
| {"id": 3, "name": "冷蔵庫", "price": 18359, "Conditions": ["NEW", 2015]}    |
+-----------------------------------------------------------------------------+

[NEW57]> update T_JSON_DOC set body = JSON_REPLACE(body,"$.price",FLOOR(10000 + (RAND() * 9000))) where id = 3;


/**** JSON_ARRAY_INSERT (SELECTでの確認とUPDATEによる更新)****/

[NEW57]> select * from T_JSON_DOC;
+----+------------------------------------------------------------------------------+------------+
| id | body                                                                         | price_only |
+----+------------------------------------------------------------------------------+------------+
|  1 | {"id": 1, "name": "自転車", "price": 10000, "Conditions": ["NEW", 2015]}     |   10000.00 |
|  2 | {"id": 2, "name": "テレビ", "price": 30000, "Conditions": ["USED",2013]}     |   30000.00 |
|  3 | {"id": 3, "name": "冷蔵庫", "price": 11154, "Conditions": ["NEW", 2015]}     |   11154.00 |
|  4 | {"id": 4, "name": "冷蔵庫", "price": 50000, "Conditions": ["NEW", 2015]}     |   50000.00 |
|  5 | {"id": 5, "name": "自転車", "price": 25000, "Conditions": ["NEW", 2015]}     |   25000.00 |
+----+------------------------------------------------------------------------------+------------+

[NEW57]>SELECT JSON_ARRAY_INSERT(body,'$.Conditions[2]','January') from T_JSON_DOC where id = 5;
+----------------------------------------------------------------------------------------+
| JSON_ARRAY_INSERT(body,'$.Conditions[2]','January')                                    |
+----------------------------------------------------------------------------------------+
| {"id": 5, "name": "自転車", "price": 25000, "Conditions": ["NEW", 2015, "January"]}    |
+----------------------------------------------------------------------------------------+

[NEW57]> update T_JSON_DOC set body = JSON_ARRAY_INSERT(body,'$.Conditions[2]','January') where id = 5;



/**** JSON_ARRAY_APPEND (SELECTでの確認とUPDATEによる更新)****/

[NEW57]> select * from T_JSON_DOC;
+----+------------------------------------------------------------------------------+------------+
| id | body                                                                         | price_only |
+----+------------------------------------------------------------------------------+------------+
|  1 | {"id": 1, "name": "自転車", "price": 10000, "Conditions": ["NEW", 2015]}     |   10000.00 |
|  2 | {"id": 2, "name": "テレビ", "price": 30000, "Conditions": ["USED",2013]}     |   30000.00 |
|  3 | {"id": 3, "name": "冷蔵庫", "price": 11154, "Conditions": ["NEW", 2015]}     |   11154.00 |
|  4 | {"id": 4, "name": "冷蔵庫", "price": 50000, "Conditions": ["NEW", 2015]}     |   50000.00 |
|  5 | {"id": 5, "name": "自転車", "price": 25000, "Conditions": ["NEW", 2015]}     |   25000.00 |
+----+------------------------------------------------------------------------------+------------+

[NEW57]> select JSON_ARRAY_APPEND(body,'$.Conditions','故障') from NEW57.T_JSON_DOC where id = 2;
+----------------------------------------------------------------------------------------+
| JSON_ARRAY_APPEND(body,'$.Conditions','故障')                                          |
+----------------------------------------------------------------------------------------+
| {"id": 2, "name": "テレビ", "price": 30000, "Conditions": ["USED", 2013, "故障"]}      |
+----------------------------------------------------------------------------------------+

[NEW57]> update T_JSON_DOC set body = JSON_ARRAY_APPEND(body,'$.Conditions','故障') where id = 2;


/**** JSON_MERGE (SELECTでの確認とUPDATEによる更新)****/

[NEW57]> select * from T_JSON_DOC;
+----+------------------------------------------------------------------------------+------------+
| id | body                                                                         | price_only |
+----+------------------------------------------------------------------------------+------------+
|  1 | {"id": 1, "name": "自転車", "price": 10000, "Conditions": ["NEW", 2015]}     |   10000.00 |
|  2 | {"id": 2, "name": "テレビ", "price": 30000, "Conditions": ["USED",2013]}     |   30000.00 |
|  3 | {"id": 3, "name": "冷蔵庫", "price": 11154, "Conditions": ["NEW", 2015]}     |   11154.00 |
|  4 | {"id": 4, "name": "冷蔵庫", "price": 50000, "Conditions": ["NEW", 2015]}     |   50000.00 |
|  5 | {"id": 5, "name": "自転車", "price": 25000, "Conditions": ["NEW", 2015]}     |   25000.00 |
+----+------------------------------------------------------------------------------+------------+

[NEW57]> SELECT JSON_MERGE(body,'{"Conditions":"Excellent"}') from T_JSON_DOC where id = 1;
+------------------------------------------------------------------------------------------+
| JSON_MERGE(body,'{"Conditions":"Excellent"}')                                            |
+------------------------------------------------------------------------------------------+
| {"id": 1, "name": "自転車", "price": 10000, "Conditions": ["NEW", 2015, "Excellent"]}    |
+------------------------------------------------------------------------------------------+


[NEW57]> update T_JSON_DOC set body = JSON_MERGE(body,'{"Conditions":"Excellent"}') where id = 1;

補足情報
json_extractと->は同等

`feature_type` varchar(30) GENERATED ALWAYS AS (json_unquote(feature->"$.type")) VIRTUAL
`feature_type` varchar(30) GENERATED ALWAYS AS (json_unquote(json_extract(`feature`,'$.type'))) VIRTUAL

メモ：
-> /* JSON_EXTRACT() */
->> /* JSON_UNQUOTE(JSON_EXTRACT()) */

必要に応じてjson_unquoteでQuoteを外して列を作成

`feature_type` varchar(30) GENERATED ALWAYS AS (json_unquote(feature->"$.type")) VIRTUAL
`feature_street` varchar(30) GENERATED ALWAYS AS (json_extract(`feature`,'$.properties.STREET')) VIRTUAL

[NEW57]> select feature_type,feature_street from T_JSON limit 10,5;
+--------------+----------------+
| feature_type | feature_street |
+--------------+----------------+
| Feature      | "JEFFERSON"    |
| Feature      | "TAYLOR"       |
| Feature      | "BEACH"        |
| Feature      | "BEACH"        |
| Feature      | "JEFFERSON"    |
+--------------+----------------+

生成列のVirtualとStoredの違いはありますが、Virtualであれば作成時も参照、更新もオンライン処理が可能。
但し、参照時にCPUを利用するのでStoredを選択するかは状況により使い分け。
Storedは実データを含む為、高速だが追加でディスク容量が必用。また、作成時は参照のみが可能になります。

ALTER TABLE T_ONLINE ALGORITHM=INPLACE, ADD feature_type varchar(30) AS (feature->"$.type") VIRTUAL;
Query OK, 0 rows affected (0.08 sec)
Records: 0  Duplicates: 0  Warnings: 0

ALTER TABLE T_ONLINE ALGORITHM=INPLACE, ADD feature_type varchar(30) AS (feature->"$.type") STORED;
ERROR 1845 (0A000): ALGORITHM=INPLACE is not supported for this operation. Try ALGORITHM=COPY.

ALTER TABLE T_ONLINE ADD feature_type varchar(30) AS (feature->"$.type") STORED;
Query OK, 206560 rows affected (6.17 sec)
Records: 206560  Duplicates: 0  Warnings: 0