Question

我有一个应用程序（将PostgreSQL 9.6迁移到10上），我想在其中检索表中的结果，并且与此时间匹配该表中存储的开放时间。

让我们用一个虚拟的例子来解释：我有一张商店表：

store_name       |  opening_hours
-----------------+-----------------------------
storeA           | ((wday between 1 and 5) and (hour between 10 and 20))
storeB           | ((wday between 2 and 5) and (hour between 9 and 18)) OR (wday in (6,7) and (hour between 9 and 12))

我想查询该表并从查询时开始抓取已打开的商店（无时区困扰）。（谁在乎谁：在我国，一周的第一天是星期一，但在此示例中我们不在乎）：

如果我的查询要求在星期三19点（晚上7点）打开商店，它将仅返回storeA。
如果查询在午夜启动，则不会选择
如果查询发生在星期四11h（11AM），则会选择两个商店。...

您能帮我使这件小事起作用吗？我认为我只是缺少正确的编写方法。

编辑：“开放时间”仅是记录我要解决此问题的方式的事情。绝不会在该数据库中添加一些新表。此处搜索的唯一答案是一种评估表中存储的表达式的方法。

Answer 1

由于您愿意接受建议，因此建议您查看该问题的可接受答案：Best way to store working hours and query it efficiently

您当前的表结构将很难管理。如果您更改了表的结构以匹配上面接受的答案，那么正是您所需要的，这将使查询非常简单。

编辑：为完整起见，链接中建议的表结构为：

要存储正常运行时间，您需要存储一些包含以下内容的记录：

商店-INTEGER
DayOfWeek-整数（0-6）
OpenTime-TIME
CloseTime-TIME

Answer 2

我对下面的回答并不完全满意，但是它可以按我想要的方式工作，而不是mysql低技术含量的方式。我在下面的工作基于How to execute a string result of a stored procedure in postgres。

如果有帮助，这里是：

-- push message to debug, to 'RAISE' usefull things
SET client_min_messages TO DEBUG;
\set VERBOSITY terse

-- must return a SETOF to evaluate my test (see RETURN QUERY EXECUTE below)
-- so here is a dirty simple [temporary] table.
CREATE TEMP TABLE stupid_bool_table (opened BOOLEAN);
INSERT INTO stupid_bool_table VALUES (true),(false);

CREATE OR REPLACE FUNCTION grab_worker_test_opening_hour(shopNametext)
  RETURNS SETOF stupid_bool_table AS
$BODY$
  DECLARE
  -- $Id: batch_workers.psql,v 1.15 2018/07/25 08:08:49 calyopea Exp $
    openhour text;
  BEGIN

  --TODO: materialized view refreshed each hours or halfs OR clever query
  SELECT INTO openhour description
         FROM shop_flat_table
        WHERE shop_id IN (select id from workers where shop=shopName)
          AND flat_txt='openhour';

  IF ( NOT FOUND ) THEN

    RAISE DEBUG 'opening_hour for % is null',shopName;
    RETURN QUERY EXECUTE 'SELECT opened FROM stupid_bool_table WHERE opened=true';   -- by DEFAULT
    -- RAISE EXCEPTION 'cant be here';   -- could be !

  ELSE

    RAISE DEBUG 'opening_hour for % is % (before replace)',shopName,openhour;

    openhour:=REPLACE(openhour,'dow', extract(dow  from NOW())::text);
    openhour:=REPLACE(openhour,'hour',extract(hour from NOW())::text);

    RAISE DEBUG 'opening_hour for % is % (after replace)',shopName,openhour;

    RETURN QUERY EXECUTE 'SELECT opened FROM stupid_bool_table WHERE opened=' || openhour;

  END IF;

END;
$BODY$
LANGUAGE plpgsql IMMUTABLE
COST 100;

所以现在：与数据：

shop  | opening_hours
------+------------------------------------------------------
ShopA | ((dow between 1 and 5) and (hour between 9 and 16)))
ShowB | ((dow between 1 and 5) and (hour between 9 and 17)))

SELECT * FROM grab_worker_test_opening_hour('ShopB');
psql:batch_workers.psql:124: DEBUG:  opening_hour for ShopB is ((dow between 1 and 5) and (hour between 9 and 17)) OR (dow in (6,7)) (before replace)
psql:batch_workers.psql:124: DEBUG:  opening_hour for ShopB is ((3 between 1 and 5) and (17 between 9 and 17)) OR (3 in (6,7)) (after replace)
 opened
--------
 t
(1 ligne)

（并在同一时间为shopA打开了= f：2018-07-25 17:15:00（等时））。

PostgreSQL PL / pgSQL：查询存储在表中（营业时间）

2 个答案: