From: Jian Shu, author: Wangcai doesn't cry

Link: https://www.jianshu.com/p/d4b012769a3b

After a long test, I will migrate to mysql8.0 as a whole; Mysql8.0 adds / optimizes many related JSON API operations for JSON operations; After reading the official documents, although most JSON operations are completed by the application layer, you can some JSON syntax of Mysql to facilitate debug ging; Select the basic and valuable parts for future reference;

https://dev.mysql.com/doc/refman/8.0/en/json.html
https://dev.mysql.com/doc/refman/8.0/en/json-utility-functions.html

Brief overview

• Null is not allowed; JSON format definition is similar to LONGBLOB or LONGTEXT; Its maximum length is max_ allowed_ Controlled by packet;

• When viewing the function of the amount of space occupied by the JSON field_ STORAGE_ SIZE(xxx);

• In addition to ordinary Json operations, some related operations of GeoJSON (geospatial data exchange format based on geometry) are supported;

• Support index for Json field (combined with the new features of Mysql8.0, the function index);

• An optional optimization item that can support partial in-situ update of Json Column is added to MySQL 8.0; The functions you can use are JSON_SET(),   JSON_REPLACE()  , JSON_REMOVE(); When using, there are some constraints, but it will have more performance;

• JSON basic tools;

  //Using JSON_ARRAY method defines JSON array;
  SELECT JSON_ARRAY(1, "abc", NULL, TRUE, CURTIME())
  //Result: [1, "abc", null, true, "11:30:24.000000"]  
  
  //JSON_ The object method defines a JSON object
  SELECT JSON_OBJECT('id', 87, 'name', 'carrot')
  //Result {"id": 87, "name": "carrot"}
  
  //The scene where arrays and objects are nested;
  [99, {"id": "HK500", "cost": 75.99}, ["hot", "cold"]] {"k1": "value", "k2": [10, 20]}
  
  //Date / time type definition
  ["12:18:29.000000", "2015-07-29", "2015-07-29 12:18:29.000000"]
  
   //JSON_QUOTE escapes the JSON object into a String, that is, escapes the internal symbols and wraps them in double quotation marks as a whole;
  JSON_QUOTE(' "null" ')
  //Result '\ "null \" "
  
  //Beautify and output JSON content;
  JSON_PRETTY()
  
  //You can convert elements inside JSON/JSON into other data types;
  //As follows, convert the id element in JSON jdoc into unsigned int;
  [https://dev.mysql.com/doc/refman/8.0/en/json.html#json-converting-between-types]    (https://dev.mysql.com/doc/refman/8.0/en/json.html#json-converting-between-types)
  ORDER BY CAST(JSON_EXTRACT(jdoc, '$.id') AS   UNSIGNED);
  
  

• Merge JSON operations   JSON_MERGE_PRESERVE() and JSON_MERGE_PATCH()   There is little possibility of actual business use;

• ->-- > operator, find the value according to the key; The difference is that -- > will remove the "of the package" and escape symbols; Its equivalent Function form is JSON_EXTRACT()

  // {"mascot": "Our mascot is a dolphin named \"Sakila\"."}
  mysql> SELECT col->"$.mascot" FROM qtest;
  //Result: "Our mascot is a dolphin named \"Sakila \ "|
  SELECT sentence->>"$.mascot" FROM facts;
  // Result: | Our mascot is a dolphin named "Sakila"|
  

• JSON Path expression
  The content in the double quotation marks after -- > above is the so-called JSON Path expression;
  This syntax is part of ECMAScript specification, so front-end programmers should be particularly familiar with it;

Take the following JSON as an example;
        [3, {"a": [5, 6], "b": 10}, [99, 100]]
        $[0] = 3 ;
        $[1] = {"a": [5, 6], "b": 10};
        $[2] = [99, 100];
         At the same time, $[1] and $[2] are not scalars, and further
        $[1].a = [5,6]
        $[1].a[1] = 6
        $[1].b = 10;
        $[2][0] = 99;
         Further supported syntax features $[n to m]
        $[ 1 to 2] = [{"a": [5, 6], "b": 10}, [99, 100]]
        $[last-2 to last-1] = [3, {"a": [5, 6], "b": 10}]
         Summarize;
         a. represents all members in object;
         b [] represents all cells in array;
         c [prefix] ** suffix represents all paths starting with prefix and ending with suffix;

• Find and modify JSON

//As above, it should be possible to use -- > syntax instead;
mysql> SELECT JSON_EXTRACT('{"a": 1, "b": 2, "c": [3, 4, 5]}', '$.*');
//[1, 2, [3, 4, 5]]  
SELECT JSON_EXTRACT('{"a": 1, "b": 2, "c": [3, 4, 5]}', '$.c[*]')
//[3, 4, 5]
SELECT JSON_EXTRACT('{"a": {"b": 1}, "c": {"b": 2}}', '$**.b');
//[1, 2]
SELECT JSON_EXTRACT('[1, 2, 3, 4, 5]', '$[1 to 3]');
//[2, 3, 4]

//JSON_SET JSON_INSERT JSON_REPLACE JSON_REMOVE
SET @j = '["a", {"b": [true, false]}, [10, 20]]';
SELECT JSON_SET(@j, '$[1].b[0]', 1, '$[2][2]', 2);
//| ["a", {"b": [1, false]}, [10, 20, 2]]    

SELECT JSON_INSERT(@j, '$[1].b[0]', 1, '$[2][2]', 2);
//["a", {"b": [true, false]}, [10, 20, 2]]

JSON_REPLACE(@j, '$[1].b[0]', 1, '$[2][2]', 2)
//["a", {"b": [1, false]}, [10, 20]]

SELECT JSON_REMOVE(@j, '$[2]', '$[1].b[1]', '$[1].b[1]');
//["a", {"b": [true]}]

• JSON Table Functions   A common scenario is that JSON data itself is a table structure;
  JSON_TABLE(*expr*, *path* COLUMNS (*column_list*) [AS] *alias*)

SELECT *  FROM JSON_TABLE(   '[{"a":"3"},{"a":2},{"b":1},{"a":0},{"a":[1,2]}]',
  ->     "$[*]"
  ->     COLUMNS(
  ->       rowid FOR ORDINALITY,
  ->       ac VARCHAR(100) PATH "$.a" DEFAULT '111' ON EMPTY DEFAULT '999' ON ERROR,
  ->       aj JSON PATH "$.a" DEFAULT '{"x": 333}' ON EMPTY,
  ->       bx INT EXISTS PATH "$.b"
  ->     )
  ->   ) AS tt;

• Comparison and Ordering of JSON Values
  I don't think it's worth it at present;

• Aggregation of JSON Values
  I don't think it's worth it at present; Aggregate functions can be used by converting the return value to other types;