在不同数量的列上连接表

Question

我想用T1中的值替换T1中属性列中的值。如果T2中的sub_id为NULL，则对于ID为T2的所有行，将T1中的 f 值替换为 t ，否则仅对具有sub_id的行执行此操作。

下面是虚拟表格和我得到的结果。

    WITH T1 AS (  SELECT 1 AS id, 2 as sub_id, '2017-08-20' AS date, 'f' AS attribute UNION ALL
          SELECT 1, 3 as sub_id, '2017-08-19', 'f' UNION ALL
          SELECT 1, 4 as sub_id,'2017-08-18', 'f' UNION ALL
          SELECT 2, 5 as sub_id,'2017-08-20', 'f' UNION ALL
          SELECT 2, 6 as sub_id,'2017-08-19', 'f' 
        ),
        T2 AS (
          SELECT 1 AS id, null as sub_id, '2017-08-19' AS date, '2017-08-25' AS date_end, 't' AS attribute UNION ALL
          SELECT 2, 5 as sub_id,'2017-08-19', '2017-08-25' AS date_end, 't' UNION ALL
          SELECT 2, 6 as sub_id,'2017-08-19', '2017-08-25' AS date_end, 't'
        )
        SELECT
          a.id,
          a.date,
          a.sub_id,
          COALESCE(b.attribute, a.attribute) as attribute
        FROM T1 AS a
        LEFT JOIN T2 AS b
        ON a.id = b.id AND a.date >= b.date AND a.date <= b.date_end 
    AND a.sub_id = case when b.sub_id is not null then a.sub_id end 

在上面的结果中，我想在所有符合条件的行中使用 t 值的属性列。 我用另一个子查询实现了我想要的东西：

WITH T1 AS (  SELECT 1 AS id, 2 as sub_id, '2017-08-20' AS date, 'f' AS attribute UNION ALL
      SELECT 1, 3 as sub_id, '2017-08-19', 'f' UNION ALL
      SELECT 1, 4 as sub_id,'2017-08-18', 'f' UNION ALL
      SELECT 2, 5 as sub_id,'2017-08-20', 'f' UNION ALL
      SELECT 2, 6 as sub_id,'2017-08-19', 'f' 
    ),
    T2 AS (
      SELECT 1 AS id, null as sub_id, '2017-08-19' AS date, '2017-08-25' AS date_end, 't' AS attribute UNION ALL
      SELECT 2, 5 as sub_id,'2017-08-19', '2017-08-25' AS date_end, 't' UNION ALL
      SELECT 2, 6 as sub_id,'2017-08-19', '2017-08-25' AS date_end, 't'
    ) , 
   adv_level AS ( SELECT
      a.id,
      a.date,
      a.sub_id,
      COALESCE(b.attribute, a.attribute) as attribute
    FROM T1 AS a
    LEFT JOIN T2 AS b
    ON a.id = b.id AND a.date >= b.date AND a.date <= b.date_end AND a.sub_id = case when b.sub_id is not null then a.sub_id end)

SELECT
      a.id,
      a.date,
      a.sub_id,
      COALESCE(b.attribute, a.attribute) as attribute
    FROM adv_level AS a
    LEFT JOIN T2 AS b
    ON a.id = b.id AND a.date >= b.date AND a.date <= b.date_end

但有没有办法只做一次？