Question

我遇到的问题是，我有一张包含数百万小时记录的表，需要创建一个查询来压缩变更数据。

示例数据 -

Start DateTime      End DateTime        Value

2015-09-13 6:00     2015-09-13 7:00     45

2015-09-13 7:00     2015-09-13 8:00     45

2015-09-13 8:00     2015-09-13 9:00     45

2015-09-13 9:00     2015-09-13 10:00    45

2015-09-13 10:00    2015-09-13 11:00    45

2015-09-13 11:00    2015-09-13 12:00    45

2015-09-13 12:00    2015-09-13 13:00    25

2015-09-13 13:00    2015-09-13 14:00    25

2015-09-13 14:00    2015-09-13 15:00    25

2015-09-13 15:00    2015-09-13 16:00    45

2015-09-13 16:00    2015-09-13 17:00    45

2015-09-13 17:00    2015-09-13 18:00    45

2015-09-13 18:00    2015-09-13 19:00    45

预期结果 -

Start DateTime      End DateTime        Value

2015-09-13 6:00     2015-09-13 12:00    45

2015-09-13 12:00    2015-09-13 15:00    25

2015-09-13 15:00    2015-09-13 19:00    45

非常感谢任何帮助。

谢谢和问候

Answer 1

排名功能可能会对您有所帮助。

假设SampleData - 是包含这些数据的表，

这可能很慢，

WITH HourlyData AS (
SELECT RANK() OVER (ORDER BY StartDate) AS RankId
    , StartDate, EndDate, Value
FROM SampleDate),

DateHourChanged AS (
SELECT DENSE_RANK () OVER  (ORDER BY (CASE WHEN A.Value <> B.value THEN 1 ELSE A.RankId  END)) AS ID
, A.StartDate, A.EndDate, A.Value
FROM HourlyData AS A
LEFT JOIN HourlyData AS B
ON A.RankId -1 = B.RankId
)

SELECT StartDate, EndDate, Value
FROM DateHourChanged
WHERE Id = 1
ORDER BY 1

我选择了上一小时和当前小时，当前小时的值与前一小时不同，然后记录当前小时。

查询将每小时数据压缩到更改

1 个答案: