如何在我的select语句中基于其他3个列创建新列

Question

我想在我的sql语句中创建一个新列，如果MRN和TF_Group被重复但PathID不同（请参见下面显示我的输出的链接），则该标志的标记为“是”

我已经能够在excel中创建该列（请参阅下面的链接，尤其是标题为“ Flag”的列），但只是想不出如何在sql server中复制它。我用来创建“标志”列的函数是= if（and（b2 = b3，d2 = d3，a2 <> a3），“是”，“否”）

我只是不知道如何在sql server中重新创建类似的东西。下面是我用来提取数据的sql代码，您可以在本文底部的链接中看到该数据。

有人有什么主意吗？

select DISTINCT 
[Pathway ID] PathID,
MRN,
DENSE_RANK() OVER(PARTITION BY mrn ORDER BY [Pathway ID]) RK,
[Treatment Function Group] TF_GROUP,
'' [Flag]

from [HRS_RTT].[dbo].[tbl_PMM_IncompletePTL_CG_Snapshot] 
where convert(date,censusdate) between '03-jun-19' and '09-jun-19'
AND MRN IS NOT NULL
AND [Treatment Function Group] IS NOT NULL

order by
MRN,
[Pathway ID]

Answer 1

如评论中所建议：

select DISTINCT 
[Pathway ID] PathID,
MRN,
CASE WHEN MRN = MRN AND [Treatment Function Group] = [Treatment Function Group] AND [Pathway ID] <> [Pathway ID] THEN 'YES' ELSE 'NO' END [Flag]
from [HRS_RTT].[dbo].[tbl_PMM_IncompletePTL_CG_Snapshot] 
where convert(date,censusdate) between '03-jun-19' and '09-jun-19'
AND MRN IS NOT NULL
AND [Treatment Function Group] IS NOT NULL

order by
MRN,
[Pathway ID]

Answer 2

这里是您所描述问题的解决方案。它使用T-SQL LEAD函数在各行之间进行比较。请注意，由于您不清楚要执行的操作尚不清楚，因此我没有在原始查询中包含DISTINCT条条件。

但是，正如我在评论中说的那样，我认为这不一定是您要执行的操作。您可能需要进一步考虑自己的要求，看看是否可以提出一种更好的方法来满足要求。

select
    PathID,
    MRN,
    TF_Group,
    case
        when MRN = Next_MRN and TF_Group = Next_TF_Group and PathID <> Next_PathID then 'Yes'
        else 'No'
    end as Flag
from (
    select
        [Pathway Id] as PathID,
        MRN,
        [Treatment Function Group] as TF_Group,

        lead([Pathway Id]) over (order by MRN, [Pathway ID]) as Next_PathID,
        lead(MRN) over (order by MRN, [Pathway ID]) as Next_MRN,
        lead([Treatment Function Group])  over (order by MRN, [Pathway ID]) as Next_TF_Group
    from tbl_PMM_IncompletePTL_CG_Snapshot
) x
order by MRN, PathID