http://www.actionsky.com/docs/archives/156
JSON (JavaScriptObject Notation) 是一种轻量级的数据交换格式,主要用于传送数据。JSON采用了独立于语言的文本格式,类似XML,但是比XML简单,易读并且易编写。对机器来说易于解析和生成,并且会减少网络带宽的传输。由于JSON格式可以解耦javascript客户端应用与Restful服务器端的方法调用,因而在互联网应用中被大量使用。
JSON的格式非常简单:名称/键值。之前MySQL版本里面要实现这样的存储,要么用VARCHAR要么用TEXT大文本。 MySQL5.7发布后,专门设计了JSON数据类型以及关于这种类型的检索以及其他函数解析。我们先看看MySQL老版本的JSON存取。
示例表结构:
CREATE TABLE json_test(
id INT,
person_desc TEXT
)ENGINE INNODB;
我们来插入一条记录:
INSERT INTO json_test VALUES (1,'{
"programmers": [{
"firstName": "Brett",
"lastName": "McLaughlin",
"email": "aaaa"
}, {
"firstName": "Jason",
"lastName": "Hunter",
"email": "bbbb"
}, {
"firstName": "Elliotte",
"lastName": "Harold",
"email": "cccc"
}],
"authors": [{
"firstName": "Isaac",
"lastName": "Asimov",
"genre": "sciencefiction"
}, {
"firstName": "Tad",
"lastName": "Williams",
"genre":"fantasy"
}, {
"firstName": "Frank",
"lastName": "Peretti",
"genre": "christianfiction"
}],
"musicians": [{
"firstName": "Eric",
"lastName": "Clapton",
"instrument": "guitar"
}, {
"firstName": "Sergei",
"lastName": "Rachmaninoff",
"instrument": "piano"
}]
}');
那一般我们遇到这样来存储JSON格式的话,只能把这条记录取出来交个应用程序,由应用程序来解析。如此一来,JSON又和特定的应用程序耦合在一起,其便利性的优势大打折扣。
现在到了MySQL5.7,可以支持对JSON进行属性的解析,我们重新修改下表结构:
ALTER TABLE json_test MODIFY person_desc json;
先看看插入的这行JSON数据有哪些KEY:
mysql> SELECT id,json_keys(person_desc) as "keys" FROM json_testG
*************************** 1. row***************************
id: 1
keys: ["authors", "musicians","programmers"]
1 row in set (0.00 sec)
我们可以看到,里面有三个KEY,分别为authors,musicians,programmers。那现在找一个KEY把对应的值拿出来:
mysql> SELECT json_extract(AUTHORS,'$.lastName[0]') AS 'name', AUTHORS FROM
-> (
-> SELECT id,json_extract(person_desc,'$.authors[0][0]') AS "authors" FROM json_test
->UNION ALL
-> SELECT id,json_extract(person_desc,'$.authors[1][0]') AS "authors" FROM json_test
-> UNION ALL
-> SELECT id,json_extract(person_desc,'$.authors[2][0]') AS "authors" FROM json_test
-> ) AS T1
-> ORDER BY NAME DESCG
*************************** 1. row***************************
name:"Williams"
AUTHORS: {"genre": "fantasy","lastName": "Williams", "firstName":"Tad"}
*************************** 2. row***************************
name:"Peretti"
AUTHORS: {"genre":"christianfiction", "lastName": "Peretti","firstName": "Frank"}
*************************** 3. row***************************
name:"Asimov"
AUTHORS: {"genre": "sciencefiction","lastName": "Asimov", "firstName":"Isaac"}
3 rows in set (0.00 sec)
现在来把详细的值罗列出来:
mysql> SELECT
->json_extract(AUTHORS,'$.firstName[0]') AS "firstname",
-> json_extract(AUTHORS,'$.lastName[0]')AS "lastname",
-> json_extract(AUTHORS,'$.genre[0]') AS"genre"
-> FROM
-> (
-> SELECT id,json_extract(person_desc,'$.authors[0]')AS "authors" FROM json
_test
-> ) AS TG
*************************** 1. row***************************
firstname: "Isaac"
lastname:"Asimov"
genre:"sciencefiction"
1 row in set (0.00 sec)
我们进一步来演示把authors 这个KEY对应的所有对象删掉。
mysql> UPDATE json_test
-> SET person_desc =json_remove(person_desc,'$.authors')G
Query OK, 1 row affected (0.01 sec)
Rows matched: 1 Changed: 1 Warnings: 0
查找下对应的KEY,发现已经被删除掉了。
mysql> SELECT json_contains_path(person_desc,'all','$.authors')as authors_exists FROM json_testG
*************************** 1. row***************************
authors_exists: 0
1 row in set (0.00 sec)
总结下,虽然MySQL5.7开始支持JSON数据类型,但是我建议如果要使用的话,最好是把这样的值取出来,然后在应用程序段来计算。毕竟数据库是用来处理结构化数据的,大量的未预先定义schema的json解析,会拖累数据库的性能。