日期具有多个天数的差异返回负数

我有一个创建的视图，我希望在交易日期中显示天数的增量。这里是我的表+ SQL语句的某些行：日期具有多个天数的差异返回负数

SELECT 
    FirstAppearenceDate 
,TranDate 
,CountryName 
,AvgTran 
,RunningTotal 
,ISNULL(RunningTotal - LAG(RunningTotal, 1) 
    OVER (PARTITION BY CountryName ORDER BY CountryName, TranDate),0) AS MoneyDelta 

,ISNULL(DATEDIFF(day,LAG(CONVERT(DATETIME, TranDate, 103), 1) OVER 
    (ORDER BY CountryName, TranDate), 
    CONVERT(DATETIME, TranDate, 103)),0) AS DaysDelta 

FROM LTV_Preprocessing_2 
ORDER BY CountryName, TranDate

，其结果是：

[前30行]：

FirstAppearenceDate| TranDate| CountryName| AvgTran| RunningTotal| MoneyDelta | DaysDelta 
20150510 20150510 Afghanistan 50.0000000000 50.0000000000 0.0000000000 0 
20150510 20150514 Afghanistan 22.0000000000 72.0000000000 22.0000000000 4 
20150510 20150603 Afghanistan 40.0000000000 112.0000000000 40.0000000000 20 
20150609 20150609 Afghanistan 25.0000000000 137.0000000000 25.0000000000 6 
20150718 20150718 Afghanistan 27.5500000000 164.5500000000 27.5500000000 39 
20150718 20150727 Afghanistan 135.6500000000 300.2000000000 135.6500000000 9 
20151026 20151026 Afghanistan 10.0000000000 310.2000000000 10.0000000000 91 
20151214 20151214 Afghanistan 20.0000000000 330.2000000000 20.0000000000 49 
20150510 20160102 Afghanistan 10.0000000000 340.2000000000 10.0000000000 19 
20150718 20160122 Afghanistan 18.5000000000 358.7000000000 18.5000000000 20 
20150510 20160616 Afghanistan 47.0000000000 405.7000000000 47.0000000000 146 
20140806 20140806 Aland Islands 10.0000000000 10.0000000000 0.0000000000 -680 
20140808 20140808 Aland Islands 20.0000000000 30.0000000000 20.0000000000 2 
20140816 20140816 Aland Islands 27.7100000000 57.7100000000 27.7100000000 8 
20140806 20140819 Aland Islands 10.0000000000 67.7100000000 10.0000000000 3 
20140820 20140820 Aland Islands 90.3300000000 158.0400000000 90.3300000000 1 
20140820 20140823 Aland Islands 15.0800000000 173.1200000000 15.0800000000 3 
20140826 20140826 Aland Islands 37.9000000000 211.0200000000 37.9000000000 3 
20140912 20140912 Aland Islands 3.0800000000 214.1000000000 3.0800000000 17 
20140919 20140919 Aland Islands 20.0000000000 234.1000000000 20.0000000000 7 
20140820 20140820 Albania 79.0400000000 79.0400000000 0.0000000000 -30 
20140821 20140821 Albania 22.0000000000 101.0400000000 22.0000000000 1 
20140821 20140822 Albania 20.0000000000 121.0400000000 20.0000000000 1 
20140821 20140824 Albania 50.0000000000 171.0400000000 50.0000000000 2 
20140821 20140827 Albania 40.0000000000 211.0400000000 40.0000000000 3 
20140821 20140828 Albania 48.0000000000 259.0400000000 48.0000000000 1 
20140831 20140831 Albania 20.0000000000 279.0400000000 20.0000000000 3 
20140821 20140901 Albania 80.0000000000 359.0400000000 80.0000000000 1 
20140821 20140902 Albania 119.0000000000 478.0400000000 119.0000000000 1 
20140821 20140910 Albania 30.0000000000 508.0400000000 30.0000000000 8

看底片，我想我的日三角洲的行为与货币三角洲完全相同 - 对于每个国家，它以0开头。在这里您可以看到它在日内达尔文栏中的阿富汗结束和奥兰群岛开始之间确实存在差异。

在此先感谢！

来源

2017-06-19 Daniel Green

你的意思是关于** **消除？，你要负的数据转换成正的数据？如果是，请使用** ABS **功能。 –

我希望这个窗口函数能像我的货币三角洲那样“理解”，对于每个国家，它应该从天数增量列中的零开始。它在第一个国家队列中工作，但之后它在一个国家的结束和下一个第一个转换日期之间有区别。 –

包括PARTITION BY国家或地区名称为daysdelta应该得到所要的结果

SELECT 
     FirstAppearenceDate 
    ,TranDate 
    ,CountryName 
    ,AvgTran 
    ,RunningTotal 
    ,ISNULL(RunningTotal - LAG(RunningTotal, 1) 
     OVER (PARTITION BY CountryName ORDER BY CountryName, TranDate),0) AS MoneyDelta 


,ISNULL(DATEDIFF(day,LAG(CONVERT(DATETIME, TranDate, 103), 1) OVER 
    (PARTITION BY CountryName ORDER BY CountryName, TranDate), 
    CONVERT(DATETIME, TranDate, 103)),0) AS DaysDelta 

FROM LTV_Preprocessing_2 
ORDER BY CountryName, TranDate

来源

2017-06-19 07:53:42

但由于某种原因，我得到了每个转账日的运行总额的重复项 - 货币delta = 0。 –

日期具有多个天数的差异返回负数

回答

相关问题