在自定义列中显示日期范围 - 差距和孤岛

Question

我有 table 看起来像这样：

+------------+------+
|    Date    | Name |
+------------+------+
| 2017-01-07 | A    |
| 2017-01-08 | A    |
| 2017-01-09 | A    |
| 2017-01-12 | A    |
| 2017-01-07 | B    |
| 2017-01-08 | B    |
| 2017-01-09 | B    |
+------------+------+

我希望能够把它变成下面这样：

+-------------------------+------+
|       Date Range        | Name |
+-------------------------+------+
| 2017-01-07 - 2017-01-09 | A    |
| 2017-01-07 - 2017-01-09 | B    |
| 2017-01-12              | A    |
+-------------------------+------+

该代码将仅查找连续日期的最小值和最大值，使用 Name 列对结果进行分组，然后将最小值和最大值列为 'to and from'一列中的字符串。

我在尝试仅列出连续日期时遇到问题。请注意，上面的第三个条目有自己的条目，因为它与较早条目中 'A' 的日期范围不连续。

编辑：请注意：这是特定于 SQL Server 2008，它不允许使用 LAG 功能。

编辑 2： McNets 提供的原始答案在 SQL Server 2012 上运行良好。我将其包含在此处，因为如果您拥有 SQL Server 2012 及更高版本，效果会更好。

;WITH CalcDiffDays AS
(
    SELECT Date, Name,
    CONCAT (Name, CAST(DATEDIFF(DAY, LAG(Date, 1, Date - 1) OVER (PARTITION BY Name ORDER BY Name, Date), Date) AS VARCHAR(10))) AS NumDays
    FROM @tmpTable
)
SELECT CONCAT(CONVERT(VARCHAR(20), MIN(Date), 102), ' - ', CONVERT(VARCHAR(20), MAX(Date), 102)) AS [Data Range], Name
FROM CalcDiffDays
GROUP BY NumDays, Name;

Answer 1

首先，我在整个 table 中添加了一个行号。

WITH RowN AS
(
    SELECT Date, Name, ROW_NUMBER() OVER (ORDER BY Name, Date) RN
    FROM #T
)

然后我加入了这个 table 本身只是为了计算日期之间的天数。

,CalcDiffDays AS
(
    SELECT RowN.Date, RowN.Name,
        ISLAND = RowN.Name + 
        CASE  
           WHEN RowN.RN > 1 AND RowN.Name = R2.Name THEN CAST(DATEDIFF(day, R2.Date, RowN.Date) AS VARCHAR(20))
           ELSE '1'
        END
    FROM RowN
         LEFT JOIN RowN R2 ON R2.RN = RowN.RN-1
)

差距。同名的连续日期之间相隔多少天。

岛屿。通过将名称添加到计算的天数。

+---------------------+------+---------+
|         Date        | Name | NumDays |
+---------------------+------+---------+
| 07.01.2017 00:00:00 |   A  |    A1   |
+---------------------+------+---------+
| 08.01.2017 00:00:00 |   A  |    A1   |
+---------------------+------+---------+
| 09.01.2017 00:00:00 |   A  |    A1   |
+---------------------+------+---------+
| 12.01.2017 00:00:00 |   A  |    A3   |
+---------------------+------+---------+
| 07.01.2017 00:00:00 |   B  |    B1   |
+---------------------+------+---------+
| 08.01.2017 00:00:00 |   B  |    B1   |
+---------------------+------+---------+
| 09.01.2017 00:00:00 |   B  |    B1   |
+---------------------+------+---------+

第二部分：获取每个岛屿的MIN和MAX Date。

WITH RowN AS
(
    SELECT Date, Name, ROW_NUMBER() OVER (ORDER BY Name, Date) RN
    FROM #T
)
,CalcDiffDays AS
(
    SELECT RowN.Date, RowN.Name,
        ISLAND = RowN.Name + 
        CASE  
           WHEN RowN.RN > 1 AND RowN.Name = R2.Name THEN CAST(DATEDIFF(day, R2.Date, RowN.Date) AS VARCHAR(20))
           ELSE '1'
        END
    FROM RowN
         LEFT JOIN RowN R2 ON R2.RN = RowN.RN-1
)    
SELECT CONVERT(VARCHAR(20), MIN(Date), 102) + ' - ' + CONVERT(VARCHAR(20), MAX(Date), 102) AS [Data Range], Name
FROM CalcDiffDays
GROUP BY ISLAND, Name
ORDER BY MIN(Date);

+-------------------------+------+
|        Data Range       | Name |
+-------------------------+------+
| 2017.01.07 - 2017.01.09 |   A  |
+-------------------------+------+
| 2017.01.07 - 2017.01.09 |   B  |
+-------------------------+------+
| 2017.01.12 - 2017.01.12 |   A  |
+-------------------------+------+

可以在这里查看：http://rextester.com/MHLEEJ50479

在自定义列中显示日期范围 - 差距和孤岛

Show Date Range in Custom Column - Gaps and Islands

sql

sql-server

sql-server-2008

tsql

gaps-and-islands