如何使重复值唯一？

Question

我在列中有重复值的记录。

例如：

INVOICENUM  PRODUCT
-------------------
  789789    ABROPT
  789789    ABROPT
  719648    AZGTEL
  754114    GRTYPT

我有2条具有重复发票编号的记录，对于第一个编号我想添加“ a”，对于第二个编号我想添加“ b”，依此类推。

INVOICENUM  PRODUCT
-------------------
 789789a    ABROPT
 789789b    ABROPT
 719648     AZGTEL
 754114     GRTYPT

如何使用SQL解决此问题-我正在使用SQL Server 2012和Microsoft SQL Server Management Studio 11.0.3000.0？

Answer 1

通常，此类问题是通过解析函数解决的。

select case
         when (count(1) over(partition by InvoiceNum) > 1) then
           InvoiceNum + Char(row_Number() over(partition by InvoiceNum order by InvoiceNum) + 96)
         else
           InvoiceNum
       end as InvoiceNum,
       Product    
  from MyTable

解析函数的主要优点是性能 。

Answer 2

好吧，假设SQL Server（具有行号功能），那么这似乎工作

select
    INVOICENUM = case 
        when INVOICENUM in (select INVOICENUM from YourTable group by INVOICENUM having COUNT(*) > 1)
        then INVOICENUM + char(96+row_number() over (partition by INVOICENUM order by INVOICENUM))
        else INVOICENUM 
    end,
    PRODUCT,
    OtherColumn,
    AnotherColumn
from 
    YourTable

虽然不是特别优雅...

这里的想法是获得一个行号，以将int转换为字符函数CHAR(...) 。 其中96是a之前字符的ASCII码-因此行号1给出a （即97的ascii码），行号2给出b （即98的ascii码），依此类推...我已经抛出了那里的case语句排除了没有重复的情况，因此它只返回InvoiceNum而无需添加字母...

这是我的测试示例...

declare @data table (INVOICENUM nvarchar(100), PRODUCT nvarchar(100))

insert into @data
values ('789789', 'ABROPT')
,('789789', 'ABROPT')
,('719648', 'AZGTEL')
,('754114', 'GRTYPT')

select
    INVOICENUM = case 
        when INVOICENUM in (select INVOICENUM from @data group by INVOICENUM having COUNT(*) > 1)
        then INVOICENUM + char(96+row_number() over (partition by INVOICENUM order by INVOICENUM))
        else INVOICENUM 
    end,
    PRODUCT
from 
    @data