如何使用 Azure 数据工厂将数据从 Excel 加载到 SQL DB

Question

我有一个 Excel 源。如何使用 Azure 数据工厂将数据从 excel 加载到 SQL DB？

Answer 1

有很多方法可以做到这一点，但我会告诉你我认为最简单的方法。

1. 在 Azure 数据工厂中设置复制数据活动以复制数据 从 excel 到 Azure SQLDB 临时表。
2. 在 Azure SQLDB 中创建一个存储过程，该过程将从暂存表插入到最终输出表中
3. 将存储过程活动连接到 ADF 中的复制数据活动，以引用您刚刚在数据库中创建的过程，这样 ADF 将在加载登台表后立即运行存储过程。
4. 您可以在加载最终输出表之后或在 ADF 中的复制活动之前使用命令/过程来清除暂存表。

请参阅下面的屏幕截图和 cmets 进行演练（在不使用 Azure 函数或其他东西的情况下，目前无法在 ADF 中以本机方式获取工作表名称，但我已尽我所能。如果 excel 是命名与可以使用的工作表相同，这是一个选项，我在这个例子中做了什么）： 整体管道视图 第一步截断 Stage Table 以确保在加载之前它是空的 第二步加载带有文件输入的 Stage 表 最后一步是运行存储过程 这是我在下面制作的存储过程的代码，以及 ADF 存储过程活动的屏幕截图：

create procedure stag.ttt_test (@ExcelFileName varchar(100))
as

DECLARE @ClassStartRow int

select @ClassStartRow = max(RowId)
from (
    select
    case when Details = '2.Class' then ROW_NUMBER() over (order by %%physloc%%) else 0 end as RowId
    from Stag.ttt_test_stage
    ) as sub
;

INSERT INTO Stag.ttt_test_final
select
    Case 
    when January is NOT NULL then DATEADD(day, -1, cast('2/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when February is NOT NULL then DATEADD(day, -1, cast('3/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when March is NOT NULL then DATEADD(day, -1, cast('4/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when April is NOT NULL then DATEADD(day, -1, cast('5/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when May is NOT NULL then DATEADD(day, -1, cast('6/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when June is NOT NULL then DATEADD(day, -1, cast('7/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when July is NOT NULL then DATEADD(day, -1, cast('8/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when August is NOT NULL then DATEADD(day, -1, cast('9/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when September is NOT NULL then DATEADD(day, -1, cast('10/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when October is NOT NULL then DATEADD(day, -1, cast('11/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when November is NOT NULL then DATEADD(day, -1, cast('12/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when December is NOT NULL then DATEADD(day, -1, cast('1/1/' + cast(YEAR(getdate()) + 1 as varchar(4)) as date))
    end as [Date],
    'name' as Category,
    Details as [Type],
    Coalesce(January, February, March, April, May, June, July, August, September, October, November, December) as [value],
    @ExcelFileName as SheetName
from (
    select
    ROW_NUMBER() over (order by %%physloc%%) as RowId,
    st.*
    from Stag.ttt_test_stage as st
    ) as sub
    where RowId > 1
    and RowId < @ClassStartRow

    ;

INSERT INTO stag.ttt_test_final
select
    Case 
    when January is NOT NULL then DATEADD(day, -1, cast('2/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when February is NOT NULL then DATEADD(day, -1, cast('3/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when March is NOT NULL then DATEADD(day, -1, cast('4/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when April is NOT NULL then DATEADD(day, -1, cast('5/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when May is NOT NULL then DATEADD(day, -1, cast('6/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when June is NOT NULL then DATEADD(day, -1, cast('7/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when July is NOT NULL then DATEADD(day, -1, cast('8/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when August is NOT NULL then DATEADD(day, -1, cast('9/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when September is NOT NULL then DATEADD(day, -1, cast('10/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when October is NOT NULL then DATEADD(day, -1, cast('11/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when November is NOT NULL then DATEADD(day, -1, cast('12/1/' + cast(YEAR(getdate()) as varchar(4)) as date))
    when December is NOT NULL then DATEADD(day, -1, cast('1/1/' + cast(YEAR(getdate()) + 1 as varchar(4)) as date))
    end as [Date],
    'class' as Category,
    Details as [Type],
    Coalesce(January, February, March, April, May, June, July, August, September, October, November, December) as [value],
    @ExcelFileName as SheetName
from (
    select
    ROW_NUMBER() over (order by %%physloc%%) as RowId,
    st.*
    from Stag.ttt_test_stage as st
    ) as sub
    where RowId > @ClassStartRow

如果这对您有用，请接受作为答案，如果您有任何其他 cmets 或问题，请告诉我。谢谢