Sql 如何分组和统计每个分组的数据数量'；s的时间范围是否在常数范围内？

我有一个包含以下字段的表：

ID（主键，整数，非空）

UpdateTime（日期时间，非空）

我试图做的是按

updatetime

对记录进行分组，并计算每个组的数量（实际上我只想得到所有组的最大计数）。对于每个组，

updateTime

的最大差值小于一个常数，例如2秒或10分钟

如何在此约束下对记录进行分组和计数

---

例如：

常数=2秒

{ID:1，更新时间：2017-03-23 00:00:00}

{ID:2，更新时间：2017-03-23 00:00:02}

{ID:3，更新时间：2017-03-23 00:00:04}

期望的结果是

第1组：{ID:1，ID:2}，第2组：{ID:2，ID:3}

所以每个组都有一个2的计数。 ID=2出现在预期的两个组中

PS：不一定要明确地找到组，我只需要（最大）计数

试试这个：

---

---

希望这对你有用

;WITH cte_inputData(Id, UpdatedTime) AS
(
SELECT 1, CAST('2017-03-23 00:00:00' AS DATETIME) UNION ALL
SELECT 2, CAST('2017-03-23 00:00:02' AS DATETIME) UNION ALL
SELECT 3, CAST('2017-03-23 00:00:04' AS DATETIME) UNION ALL
SELECT 4, CAST('2017-03-23 00:00:06' AS DATETIME) UNION ALL
SELECT 5, CAST('2017-03-23 00:00:16' AS DATETIME) UNION ALL
SELECT 6, CAST('2017-03-23 00:00:26' AS DATETIME) UNION ALL
SELECT 7, CAST('2017-03-23 00:00:36' AS DATETIME) UNION ALL
SELECT 8, CAST('2017-03-23 00:00:38' AS DATETIME) UNION ALL
SELECT 9, CAST('2017-03-23 00:00:40' AS DATETIME) UNION ALL
SELECT 10, CAST('2017-03-23 00:00:50' AS DATETIME) UNION ALL
SELECT 11, CAST('2017-03-23 00:01:00' AS DATETIME)
),
--Compensate for any Missing sequences in the ID
cte_Sequence(Id, OriginalID, UpdatedTime) AS (
    SELECT ROW_NUMBER() OVER (
            ORDER BY ID
            ) AS Id,
        Id AS OriginalID,
        UpdatedTime
    FROM cte_inputData
    ),
cte_BuildResult AS (
    SELECT b.Id,
        CAST(b.Id AS VARCHAR(10)) + ',' + CAST(a.OriginalID AS VARCHAR(10)) AS GroupNumber,
        DATEDIFF(SECOND, b.UpdatedTime, a.UpdatedTime) AS Differance
    FROM cte_Sequence a
    INNER JOIN cte_inputData b
        ON a.Id = b.ID + 1
    ),
cte_resultBuilderFinal(ID, Differance) AS (
    SELECT TOP 1 MAX(Id),
        MAX(Differance)
    FROM cte_BuildResult
    GROUP BY Differance
    ORDER BY MAX(Differance) DESC
    )

SELECT GroupNumber,
Differance
FROM cte_BuildResult b
WHERE EXISTS (
    SELECT 1
    FROM cte_resultBuilderFinal a
    WHERE a.ID = b.Id
    )

---

请编辑您的问题并提供示例数据和所需结果。而且，你的问题毫无意义。如果按

updateTime

分组，则结果集中的每一行都将有一个时间，最大值与最小值相同。@GordonLinoff更新了问题，抱歉混淆，我不是指SQL关键字“分组依据”，我想说的是如何找到

updatetime

中包含上述约束的数据组…再次为我的糟糕英语道歉

;WITH cte_inputData(Id, UpdatedTime) AS
(
SELECT 1, CAST('2017-03-23 00:00:00' AS DATETIME) UNION ALL
SELECT 2, CAST('2017-03-23 00:00:02' AS DATETIME) UNION ALL
SELECT 3, CAST('2017-03-23 00:00:04' AS DATETIME) UNION ALL
SELECT 4, CAST('2017-03-23 00:00:06' AS DATETIME) UNION ALL
SELECT 5, CAST('2017-03-23 00:00:16' AS DATETIME) UNION ALL
SELECT 6, CAST('2017-03-23 00:00:26' AS DATETIME) UNION ALL
SELECT 7, CAST('2017-03-23 00:00:36' AS DATETIME) UNION ALL
SELECT 8, CAST('2017-03-23 00:00:38' AS DATETIME) UNION ALL
SELECT 9, CAST('2017-03-23 00:00:40' AS DATETIME) UNION ALL
SELECT 10, CAST('2017-03-23 00:00:50' AS DATETIME) UNION ALL
SELECT 11, CAST('2017-03-23 00:01:00' AS DATETIME)
),
--Compensate for any Missing sequences in the ID
cte_Sequence(Id, OriginalID, UpdatedTime) AS (
    SELECT ROW_NUMBER() OVER (
            ORDER BY ID
            ) AS Id,
        Id AS OriginalID,
        UpdatedTime
    FROM cte_inputData
    ),
cte_BuildResult AS (
    SELECT b.Id,
        CAST(b.Id AS VARCHAR(10)) + ',' + CAST(a.OriginalID AS VARCHAR(10)) AS GroupNumber,
        DATEDIFF(SECOND, b.UpdatedTime, a.UpdatedTime) AS Differance
    FROM cte_Sequence a
    INNER JOIN cte_inputData b
        ON a.Id = b.ID + 1
    ),
cte_resultBuilderFinal(ID, Differance) AS (
    SELECT TOP 1 MAX(Id),
        MAX(Differance)
    FROM cte_BuildResult
    GROUP BY Differance
    ORDER BY MAX(Differance) DESC
    )

SELECT GroupNumber,
Differance
FROM cte_BuildResult b
WHERE EXISTS (
    SELECT 1
    FROM cte_resultBuilderFinal a
    WHERE a.ID = b.Id
    )