Question

我准备了sql小提琴：http://sqlfiddle.com/#!15/62e65/2

架构就在这里：

CREATE TABLE products
    ("id" int, "name" varchar(5))
;

INSERT INTO products
    ("id", "name")
VALUES
    (1, 'Car'),
    (2, 'Phone')
;

CREATE TABLE operations
    ("id" int, "product_id" int, "status" varchar(7), "type" varchar(8))
;

INSERT INTO operations
    ("id", "product_id", "status", "type")
VALUES
    (1, 1, 'pending', 'invoice'),
    (2, 1, 'done', 'delivery'),
    (3, 2, 'done', 'delivery'),
    (3, 2, 'done', 'invoice')
;

我知道数据架构可能更好，但我现在没有可能重构它 - 我只是添加新视图。关于架构的注意事项：产品总是有2个操作：开票和交货。我想要达到的目的是获得这样的结果：

name status
car pending
phone done

产品状态是检查两个产品操作后返回的字符串规则是产品状态仅在两个操作完成时完成，否则其待定如何在postgres中编写这样的查询？

Answer 1

SELECT p.name, CASE WHEN status='done' THEN 'done' else 'pending' END
FROM (
    SELECT p.id, p.name, string_agg(distinct o.status, '') as status
    FROM products p JOIN operations o ON o.product_id = p.id
    GROUP BY p.id, p.name
) sub

我们在一个字符串（string_agg）中连接不同的值，然后在主查询中检查：如果ne string是单'done'，那么所有操作都是'done'。否则，一个或两个操作都是'pending'。

由于您的外键是products.id <--> operations.product_id，我们必须GROUP BY products.id。

Answer 2

select
    p.name,
    case
        when count(case when status <> 'done' then 1 else null end) = 0 then 'done'
        else 'pending'
    end status
  from products p
  inner join operations o on p.id = o.product_id
  group by p.name, p.id;

Answer 3

您可以创建两个子查询，一个用于订单，一个用于交付，加入它们，然后使用case表达式检查两者的状态：

SELECT invoice.name, 
       CASE WHEN invoice.status = 'done' and delivery.status = 'done' THEN 'done' 
            ELSE 'pending' 
       END AS status
FROM   (SELECT p.name, p.id, o.status
        FROM   products p
        JOIN   operations o ON o.type = 'invoice' AND 
                               o.product_id = p.id) invoice
JOIN   (SELECT p.name, p.id, o.status
        FROM   products p
        JOIN   operations o ON o.type = 'delivery' AND
                               o.product_id = p.id) delivery
       ON invoice.id = delivery.id

如何在postgres中将查询结果解析为字符串？

3 个答案: