查询以获取表中的第一行和最后一行答案

【问题标题】：Query to get the first and last row in a table查询以获取表中的第一行和最后一行
【发布时间】：2016-11-15 01:00:55
【问题描述】：

我运行了以下查询以从表中获取第一行和最后一行。

SELECT Start, End, Start1, End1
FROM
(SELECT Start AS Start, End AS End 
  FROM [TB1]  WHERE Id = 251 ORDER BY DateTime DESC LIMIT 1),
(SELECT Start1 AS Start1, End1 AS End1 
  FROM [TB2]  WHERE Id = 251 ORDER BY DateTime LIMIT 1);

我得到如下输出：

 Start      End      Start1      End1
 25.32      -98.55    null       null
 null       null      29.81      -98.32

我正在尝试获取以下格式的输出：

 Start      End      Start1      End1
 25.32      -98.55   29.81       -98.32

【问题讨论】：

标签： sql google-bigquery

【解决方案1】：

我觉得你的问题是“部分的”，接下来将是如何对所有 Id 做同样的事情！见下文

SELECT Id, Start, End, Start1, End1
FROM
(SELECT Id, Start, End 
 FROM (
   SELECT
     Id, Start AS Start, End AS END, 
     ROW_NUMBER() OVER(PARTITION BY Id ORDER BY DateTime DESC) AS win
   FROM [TB1])
 WHERE win = 1
) AS t1
JOIN
(SELECT Id, Start1, End1 
 FROM (
   SELECT
     Id, Start1 AS Start1, End1 AS End1,
     ROW_NUMBER() OVER(PARTITION BY Id ORDER BY DateTime) AS win
   FROM [TB2])
 WHERE win = 1
) AS t2
ON t1.Id = t2.Id

【讨论】：

【解决方案2】：

使用max，这样结果将位于一行。

SELECT max(Start) start, max(End) end, max(Start1) start1, max(End1) end1
FROM
(SELECT Start AS Start, End AS End 
  FROM [TB1]  WHERE Id = 251 ORDER BY DateTime DESC LIMIT 1),
(SELECT Start1 AS Start1, End1 AS End1 
  FROM [TB2]  WHERE Id = 251 ORDER BY DateTime LIMIT 1);

【讨论】：

【解决方案3】：

如果您想要单个表（而不是两个）上的第一行和最后一行，您可以使用 ARRAY_AGG 和 ORDER BY 和 LIMIT 在 standard SQL 中计算它（取消选中“使用旧版 SQL”下的“在 UI 中显示选项）。

WITH T AS (
  SELECT 251 AS Id, 10.0 AS Start, -15 AS `End`, DATE('2016-11-09') AS DateTime UNION ALL
  SELECT 251, 25.32 AS Start, -98.55, DATE('2016-11-10') UNION ALL
  SELECT 251, 29.81, -98.32, DATE('2016-01-01')
)
SELECT
  ARRAY_AGG(STRUCT(Start, `End`)
            ORDER BY DateTime DESC LIMIT 1)[OFFSET(0)] AS first_row,
  ARRAY_AGG(STRUCT(Start AS Start1, `End` AS End1)
            ORDER BY DateTime ASC LIMIT 1)[OFFSET(0)] AS last_row
FROM T
WHERE Id = 251;

总的来说，这也可能比使用解析函数快一点，但我没有做过任何具体的比较。

【讨论】：