SQLAlchemy ORM - 急切加载

急切加载减少了查询次数。SQLAlchemy 提供了通过查询选项调用的急切加载函数，这些函数为 Query 提供了额外的指令。这些选项决定了如何通过 Query.options() 方法加载各种属性。

子查询加载

我们希望 Customer.invoices 能够急切加载。orm.subqueryload() 选项提供了第二个 SELECT 语句，该语句完全加载与刚刚加载的结果关联的集合。名称“subquery”导致 SELECT 语句通过 Query 重新使用并嵌入到针对相关表的 SELECT 中作为子查询来构建。

from sqlalchemy.orm import subqueryload
c1 = session.query(Customer).options(subqueryload(Customer.invoices)).filter_by(name = 'Govind Pant').one()

这将产生以下两个 SQL 表达式：

SELECT customers.id 
AS customers_id, customers.name 
AS customers_name, customers.address 
AS customers_address, customers.email 
AS customers_email
FROM customers
WHERE customers.name = ?
('Govind Pant',)

SELECT invoices.id 
AS invoices_id, invoices.custid 
AS invoices_custid, invoices.invno 
AS invoices_invno, invoices.amount 
AS invoices_amount, anon_1.customers_id 
AS anon_1_customers_id
FROM (
   SELECT customers.id 
   AS customers_id
   FROM customers
   WHERE customers.name = ?) 
   
AS anon_1 
JOIN invoices 
ON anon_1.customers_id = invoices.custid 
ORDER BY anon_1.customers_id, invoices.id 2018-06-25 18:24:47,479 
INFO sqlalchemy.engine.base.Engine ('Govind Pant',)

要访问两个表中的数据，我们可以使用以下程序：

print (c1.name, c1.address, c1.email)

for x in c1.invoices:
   print ("Invoice no : {}, Amount : {}".format(x.invno, x.amount))

上述程序的输出如下：

Govind Pant Gulmandi Aurangabad [email protected]
Invoice no : 3, Amount : 10000
Invoice no : 4, Amount : 5000

连接加载

另一个函数称为 orm.joinedload()。这会发出一个 LEFT OUTER JOIN。引导对象以及相关对象或集合在一步骤中加载。

from sqlalchemy.orm import joinedload
c1 = session.query(Customer).options(joinedload(Customer.invoices)).filter_by(name='Govind Pant').one()

这会发出以下表达式，并产生与上述相同的输出：

SELECT customers.id 
AS customers_id, customers.name 
AS customers_name, customers.address 
AS customers_address, customers.email 
AS customers_email, invoices_1.id 
AS invoices_1_id, invoices_1.custid 
AS invoices_1_custid, invoices_1.invno 
AS invoices_1_invno, invoices_1.amount 
AS invoices_1_amount

FROM customers 
LEFT OUTER JOIN invoices 
AS invoices_1 
ON customers.id = invoices_1.custid

WHERE customers.name = ? ORDER BY invoices_1.id
('Govind Pant',)

OUTER JOIN 导致了两行，但它只返回了一个 Customer 实例。这是因为 Query 基于对象标识对返回的实体应用了“唯一化”策略。连接急切加载可以在不影响查询结果的情况下应用。

subqueryload() 更适合加载相关集合，而 joinedload() 更适合多对一关系。