将不重叠数据的多行合并为一行 SQL 服务器

Question

我有一个数据 table，数据如下：

ID	Task	time
Jim	sleep	5:50
Jim	wakeup	7:15
Bob	sleep	6:00
Bob	brushteeth	8:00
Bob	eat	9:00

我是运行一个生成以下内容的查询

SELECT 
    dbo.Person.ID, 
    CAST(CASE WHEN dbo.DailyActivities.Activity = 'sleep' THEN 1 ELSE 0 END AS bit) AS slept, 
    CAST(CASE WHEN dbo.DailyActivities.Activity = 'wakeup' THEN 1 ELSE 0 END AS bit) AS wokeup, 
    CAST(CASE WHEN dbo.DailyActivities.Activity = 'brushteeth' THEN 1 ELSE 0 END AS bit) AS brushedteeth, 
    CAST(CASE WHEN dbo.DailyActivities.Activity = 'eat' THEN 1 ELSE 0 END AS bit) AS ate
FROM 
    dbo.Person 
LEFT OUTER JOIN
    dbo.DailyActivities ON dbo.Person.ID = dbo.DailyActivities.ID

ID	slept	wokeup	brushedteeth	ate
Jim	TRUE	FALSE	FALSE	FALSE
Jim	FALSE	TRUE	FALSE	FALSE
Bob	TRUE	FALSE	FALSE	FALSE
Bob	FALSE	FALSE	TRUE	FALSE
Bob	FALSE	FALSE	FALSE	TRUE

我们如何获得如下所示每行 1 个 ID 的结果

ID	slept	wokeup	brushedteeth	ate
Jim	TRUE	TRUE	FALSE	FALSE
Bob	TRUE	FALSE	TRUE	TRUE

Answer 1

最简单的方法是使用 DISTINCT + 子查询 + CASE。我稍微简化了您的查询，所以没有加入等，所以我必须少输入一些。但我希望它背后的概念是清楚的。

SELECT DISTINCT
    p.[ID]
  , CASE
      WHEN (SELECT COUNT(*) FROM [dbo].[Person] sub WHERE sub.[ID] = p.[ID] AND sub.[Task] = 'slept') > 0 THEN CAST(1 as bit)
      ELSE CAST(0 as bit)
    END AS [slept]
  , CASE
      WHEN (SELECT COUNT(*) FROM [dbo].[Person] sub WHERE sub.[ID] = p.[ID] AND sub.[Task] = 'wakeup') > 0 THEN CAST(1 as bit)
      ELSE CAST(0 as bit)
    END AS [wokeup]
  , CASE
      WHEN (SELECT COUNT(*) FROM [dbo].[Person] sub WHERE sub.[ID] = p.[ID] AND sub.[Task] = 'brushteeth') > 0 THEN CAST(1 as bit)
      ELSE CAST(0 as bit)
    END AS [brushteeth]
  , CASE
      WHEN (SELECT COUNT(*) FROM [dbo].[Person] sub WHERE sub.[ID] = p.[ID] AND sub.[Task] = 'eat') > 0 THEN CAST(1 as bit)
      ELSE CAST(0 as bit)
    END AS [ate]
FROM [dbo].[Person] p

Answer 2

正如我在评论中提到的，在 CASE 表达式周围使用 MAX；这称为条件聚合：

SELECT P.ID, 
       CAST(MAX(CASE WHEN DA.Activity = 'sleep' THEN 1 ELSE 0 END) AS bit) AS slept, 
       CAST(MAX(CASE WHEN DA.Activity = 'wakeup' THEN 1 ELSE 0 END) AS bit) AS wokeup, 
       CAST(MAX(CASE WHEN DA.Activity = 'brushteeth' THEN 1 ELSE 0 END) AS bit) AS brushedteeth, 
       CAST(MAX(CASE WHEN DA.Activity = 'eat' THEN 1 ELSE 0 END) AS bit) AS ate
FROM dbo.Person P
     LEFT OUTER JOIN dbo.DailyActivities DA ON P.ID = DA.ID
GROUP BY P.ID

此外，正如我在评论中提到的，我删除了列上的 3+ 部分命名，因为它 will be deprecated。

将不重叠数据的多行合并为一行 SQL 服务器

Merge multiple rows without overlapping data into one row SQL Server

sql

sql-server