根据扩展的日期范围连接表

Question

我有两个 table、orders 和 currency_rates。我需要使用 orders.received_at 和 currency_rates.valid_from 合并订单（并且还使用公共列 currency_id，但这很简单）。

我面临的问题是 currency_rates.valid_from 不是完整的 table 日期，而只是日期范围的开始和结束日期。实际上，我有三个货币周期 '2019-12-01' - '2020-01-03'、'2020-01-03' - '2020-01-09'、'2020-01-09' - onward。在 Postgres 中最优雅的实现方式是什么？是否可以使用 Select 语句来做到这一点？谢谢！

编辑：欢迎不手动指定范围的解决方案。

EDIT2：添加了一个 table orders_currency_rates，其中包含一些所需结果的示例行。

    CREATE TABLE orders
(
    id                  BIGINT,
    received_at         DATE,
    shipping_cost       DOUBLE PRECISION,
    currency_id         VARCHAR,
    invoice_address_id  BIGINT,
    delivery_address_id BIGINT
);

INSERT INTO orders (id, received_at, shipping_cost, currency_id, invoice_address_id, delivery_address_id)
VALUES (385902, '2020-01-01', 0, 'CZK', 1, 11),
       (386278, '2020-01-02', 12.83, 'USD', 2, NULL),
       (386279, '2020-01-03', 49.36, 'USD', 3, 12),
       (386280, '2020-01-03', 12.83, 'USD', 4, 13),
       (386281, '2020-01-05', 12.83, 'USD', 5, 14),
       (386282, '2020-01-06', 11.43, 'GBP', 6, NULL),
       (386283, '2020-01-07', 12.83, 'USD', 7, 15),
       (386284, '2020-01-08', 44.03, 'EUR', 8, NULL),
       (386285, '2020-01-11', 12.83, 'USD', 9, NULL),
       (386286, '2020-02-12', 62.55, 'USD', 10, NULL);
CREATE TABLE currency_rates
(
    currency_id VARCHAR,
    rate        DOUBLE PRECISION,
    valid_from  DATE
);

INSERT INTO currency_rates (currency_id, rate, valid_from)
VALUES ('EUR', 24.165, '2019-12-01'),
       ('USD', 19.359, '2019-12-01'),
       ('GBP', 27.039, '2019-12-01'),
       ('PLN', 5.5, '2019-12-01'),
       ('EUR', 25.2, '2020-01-03'),
       ('USD', 20.34, '2020-01-03'),
       ('GBP', 28.4, '2020-01-03'),
       ('PLN', 5.3, '2020-01-03'),
       ('EUR', 26.165, '2020-01-09'),
       ('USD', 21.359, '2020-01-09'),
       ('GBP', 29.039, '2020-01-09'),
       ('PLN', 5.8, '2020-01-09');

连接所需结果的示例。

     CREATE TABLE orders_currency_rates
(
    id                  BIGINT,
    received_at         DATE,
    shipping_cost       DOUBLE PRECISION,
    currency_id         VARCHAR,
    invoice_address_id  BIGINT,
    delivery_address_id BIGINT,
    rate                DOUBLE PRECISION,
    valid_from          DATE
);

INSERT INTO orders_currency_rates (id, received_at, shipping_cost, currency_id, invoice_address_id, delivery_address_id, rate, valid_from)
VALUES (386278, '2020-01-02', 12.83, 'USD', 2, NULL, 19.359, '2019-12-01'),
       (386279, '2020-01-03', 49.36, 'USD', 3, 12, 20.34, '2020-01-03'),
       (386286, '2020-02-12', 62.55, 'USD', 10, NULL, 21.359, '2020-01-09');

Answer 1

我会用

SELECT ...
FROM (SELECT CASE WHEN received_at >= '2020-01-09'
                  THEN DATE '2020-01-09'
                  WHEN received_at >= '2020-01-03'
                  THEN DATE '2020-01-03'
                  ELSE DATE '2019-12-01'
             END AS valid_from,
             ...
      FROM orders
     ) AS orders_adj
   JOIN currency_rates USING (valid_from);

Answer 2

如何实现这个范围查询？

关键是在每个`LEFT JOIN`.

中过滤两次数据

1. 过滤掉询问时间早于费率时间的部分。
1. 按max过滤最接近的通缉时间。

SELECT * FROM orders o LEFT JOIN currency_rates c
ON c.currency_id = o.currency_id AND c.valid_from = (
  SELECT max(valid_from) FROM currency_rates
  WHERE valid_from <= o.received_at
);

/* My result:
   id   | received_at | shipping_cost | currency_id | invoice_address_id | delivery_address_id | currency_id |  rate  | valid_from 
--------+-------------+---------------+-------------+--------------------+---------------------+-------------+--------+------------
 385902 | 2020-01-01  |             0 | CZK         |                  1 |                  11 |             |        | 
 386284 | 2020-01-08  |         44.03 | EUR         |                  8 |                     | EUR         |   25.2 | 2020-01-03
 386282 | 2020-01-06  |         11.43 | GBP         |                  6 |                     | GBP         |   28.4 | 2020-01-03
 386278 | 2020-01-02  |         12.83 | USD         |                  2 |                     | USD         | 19.359 | 2019-12-01
 386281 | 2020-01-05  |         12.83 | USD         |                  5 |                  14 | USD         |  20.34 | 2020-01-03
 386279 | 2020-01-03  |         49.36 | USD         |                  3 |                  12 | USD         |  20.34 | 2020-01-03
 386280 | 2020-01-03  |         12.83 | USD         |                  4 |                  13 | USD         |  20.34 | 2020-01-03
 386283 | 2020-01-07  |         12.83 | USD         |                  7 |                  15 | USD         |  20.34 | 2020-01-03
 386285 | 2020-01-11  |         12.83 | USD         |                  9 |                     | USD         | 21.359 | 2020-01-09
 386286 | 2020-02-12  |         62.55 | USD         |                 10 |                     | USD         | 21.359 | 2020-01-09
(10 rows)
*/

Answer 3

使用横向连接！

select o.*, cr.*
from orders o left join lateral
     (select cr.*
      from currency_rates cr
      where cr.currency_id = o.currency_id and
            cr.valid_from <= o.received_at
      order by cr.valid_from desc
      where 1=1
     ) cr
     on 1=1

根据扩展的日期范围连接表

Joining tables based on expanded date range

sql

postgresql

postgresql-13

如何实现这个范围查询？

关键是在每个`LEFT JOIN`.

根据扩展的日期范围连接表

Joining tables based on expanded date range

sql

postgresql

postgresql-13

如何实现这个范围查询？

关键是在每个LEFT JOIN.

关键是在每个`LEFT JOIN`.