Question

在以下示例中，我预期结果

Row    a    b    f0_     
1    1    1    3     
2    1    2    3     
3    1    3    5     
4    1    4    5     
5    1    5    null

因为通常聚合倾向于忽略空值。如果FIRST_VALUE不忽略空值，则使用LEAD

会有什么价值

示例：

select a, b, first_value(c) over (partition by a order by b asc rows BETWEEN 1 following AND UNBOUNDED FOLLOWING)
from 
(select 1 as a, 1 as b, 1 as c),
(select 1 as a, 2 as b, null as c),
(select 1 as a, 3 as b, 3 as c),
(select 1 as a, 4 as b, null as c),
(select 1 as a, 5 as b, 5 as c),

给出

Row    a    b    f0_     
1    1    1    null     
2    1    2    3     
3    1    3    null     
4    1    4    5     
5    1    5    5

Answer 1

我原本期待结果

下面的技巧给出了预期的（在您的问题中）结果

SELECT
  a, b, 
  MAX(c) OVER (PARTITION BY a ORDER BY grp ASC RANGE BETWEEN 1 FOLLOWING AND 1 FOLLOWING)
FROM (
  SELECT 
    a, b, c, 
    COUNT(c) OVER (PARTITION BY a ORDER BY b ASC rows BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) grp
  FROM 
    (SELECT 1 AS a, 1 AS b, 1 AS c),
    (SELECT 1 AS a, 2 AS b, NULL AS c),
    (SELECT 1 AS a, 3 AS b, 3 AS c),
    (SELECT 1 AS a, 4 AS b, NULL AS c),
    (SELECT 1 AS a, 5 AS b, 5 AS c)
)

它比使用LEAD有什么价值

LEAD有更多触及签名 - LEAD(<expr>[, <offset>[, <default_value>]]) - 所以如果你只需要第一个值，你可以将其缩短为FIRST_VALUE(<field_name>) - 我认为这是主要的实际差异

BigQuery，FIRST_VALUE和null

1 个答案: