同一列 SQL Server Management Studio 中行的 YoY 计算答案

【问题标题】：YoY Calculations on Rows in Same Column SQL Server Management Studio同一列 SQL Server Management Studio 中行的 YoY 计算
【发布时间】：2017-07-18 21:34:55
【问题描述】：

Business Unit |  Date  | Sales
North America | 5/1/17 | 5,000
North America | 4/1/17 | 4,000
North America | 3/1/17 | 3,000
North America | 2/1/17 | 2,000
North America | 1/1/17 | 1,000
          ............
North America | 5/1/16 | 4,000
North America | 4/1/16 | 3,000
North America | 3/1/16 | 2,000
North America | 2/1/16 | 2,000
North America | 1/1/16 |   500
         ............
Asia          | 5/1/17 | 5,000
Asia          | 4/1/17 | 2,000
Asia          | 3/1/17 | 1,000
Asia          | 2/1/17 | 3,000
Asia          | 1/1/17 | 1,500
          ............
Asia          | 5/1/16 | 4,000
Asia          | 4/1/16 | 2,000
Asia          | 3/1/16 |   500
Asia          | 2/1/16 | 1,500
Asia          | 1/1/16 | 1,000

每个月都由 38、30 或 31 天组成 - 但是为了方便示例，我只在上面的数据中显示了前几天。

以下是我希望实现的输出示例：

Business Unit |  Date  | Sales | Year over Year Change |
North America | 5/1/17 | 5,000 |           25%         |
North America | 4/1/17 | 4,000 |           33%         |
North America | 3/1/17 | 3,000 |           50%         |
North America | 2/1/17 | 2,000 |            0%         |
North America | 1/1/17 | 1,000 |           50%         |
          ............
North America | 5/1/16 | 4,000 |           xx%         |
North America | 4/1/16 | 3,000 |           xx%         |
North America | 3/1/16 | 2,000 |           xx%         |
North America | 2/1/16 | 2,000 |           xx%         |
North America | 1/1/16 |   500 |           xx%         |
         ............
Asia          | 5/1/17 | 5,000 |           25%         |
Asia          | 4/1/17 | 2,000 |            0%         |
Asia          | 3/1/17 | 1,000 |          100%         |
Asia          | 2/1/17 | 3,000 |          100%         |
Asia          | 1/1/17 | 1,500 |           50%         |
          ............
Asia          | 5/1/16 | 4,000 |           xx%         |
Asia          | 4/1/16 | 2,000 |           xx%         |
Asia          | 3/1/16 |   500 |           xx%         |
Asia          | 2/1/16 | 1,500 |           xx%         |
Asia          | 1/1/16 | 1,000 |           xx%         |

如上所述，这些数据是每日数据，可以追溯到多年前。非常感谢您的任何帮助！

【问题讨论】：

标签： sql-server sql-server-2012 ssms

【解决方案1】：

您可以如下查询：

Select coalesce(lastyear.[Business Unit], currentyear.[Business Unit]) as BusinessUnit, currentyear.[date], lastyear.[date] as [LastYear Date],
    Format((1-(lastyear.Sales*1.0)/currentyear.Sales), 'p') 
from #yoydata lastyear
left join #yoydata currentyear
on lastyear.[date] = dateadd(yy,-1,currentyear.[date])
and lastyear.[Business Unit] = currentyear.[Business Unit]

【讨论】：

非常感谢您非常接近。不过，实际计算列似乎已关闭。当我将它插入到我的实际数据中时，它会提取正确的日期，但是它不会计算根据去年日期变化的年份
这有意义吗？
你遇到了什么问题？
是的，你是对的，复制粘贴错误和去年。[Business Unit] = currentyear.[Business Unit]

【解决方案2】：

SELECT
    COALESCE(lastyear.[bu], currentyear.[bu]) AS BusinessUnit
    , currentyear.[dt]
    , lastyear.[dt]                             AS [LastYear Date]
    , Format((1-(lastyear.Sales*1.0)/currentyear.Sales), 'p') 
FROM
    #yoydata lastyear 
        LEFT JOIN #yoydata currentyear  ON lastyear.[dt] = dateadd(yy,-1,currentyear.[dt]) 
                                            AND lastyear.bu = currentyear.bu

【讨论】：

上一个答案只是缺少业务部门的加入。