Question

基本上我的问题是我需要删除不正确的重复值（请参见下文）。我无法使用Destinct函数，因为它会删除一些正确的值。如果您有任何建议我会很感激。如果您需要更多说明，请告诉我:)

我有两张桌子。

----------------------------------------------------------
CUSTOMER | Ammount | Invoice number | Time Stamp
----------------------------------------------------------
 A       | 57000,2 | 631            | Time Stamp
 A       | 56000   | 631            | Time Stamp
 A       | 55000,1 | 632            | Time Stamp
 A       | 54000   | 632            | Time Stamp

和

--------------------------------------------------------------------------
CUSTOMER |        FREE TEXT           |Invoice number| Time Stamp
--------------------------------------------------------------------------
 A       | 57.000,2 invoice number 631 | 631          | Time Stamp
 A       | 56.000   invoice number 631 | 631          | Time Stamp
 A       | 55.000,1 invoice number 632 | 632          | Time Stamp
 A       | 54.000   invoice number 632 | 632          | Time Stamp

我使用此查询：

Select A.CUTOMER, A.AMMOUNT, B.FREE_TEXT, B.Invoice_number   
FROM Table1 A,
Table2 B
WHERE A.CUSTOMER = B.CUSTOMER
AND A.Invoice_number = B.Invoice_number
AND B.Invoice_number IN ('631','632')
AND A.CUSTOMER = 'A'
AND B.Time_stamp >= TIMESTAMP('2015-01-01 00:00:00')
AND A.Time_stamp >= TIMESTAMP('2015-01-01 00:00:00')

结果是重复的，其中1个不正确，结果如下：

 A       | 57000,2 | 57.000,2 invoice number 631  | 631 
 A       | 56000   | 57.000,2 invoice number 631  | 631
 A       | 57000,2 | 56.000   invoice number 631  | 631 
 A       | 56000   | 56.000   invoice number 631  | 631 
 A       | 55000,1 | 55.000,1 invoice number 632  | 632 
 A       | 54000   | 54.000   invoice number 632  | 632 
 A       | 55000,1 | 55.000,1 invoice number 632  | 632 
 A       | 54000   | 54.000   invoice number 632  | 632

我希望它像：

 A       | 57000,2 | 57.000,2 invoice number 631  | 631  |
 A       | 56000   | 56.000   invoice number 631  | 631  |
 A       | 55000,1 | 55.000,1 invoice number 632  | 632  |
 A       | 54000   | 54.000   invoice number 632  | 632  |

Answer 1

按条件使用分组。

Select A.CUTOMER, A.AMMOUNT, B.FREE_TEXT, B.Invoice_number   
FROM Table1 A,
Table2 B
WHERE A.CUSTOMER = B.CUSTOMER
AND A.Invoice_number = B.Invoice_number
AND B.Invoice_number IN ('631','632')
AND A.CUSTOMER = 'A'
AND B.Time_stamp >= TIMESTAMP('2015-01-01 00:00:00')
AND A.Time_stamp >= TIMESTAMP('2015-01-01 00:00:00')
GROUP BY A.AMMOUNT

Answer 2

在这种情况下，我想了解这些表格的关系。根据您的评论，它取决于客户，invoice_number和金额，但金额在一个表格中的自由格式文本字段中。

如果我们假设格式在此free_text字段中是一致的....我们假设mySQl与Oracle。

SELECT A.CUTOMER, A.AMMOUNT, B.FREE_TEXT, B.Invoice_number   
FROM Table1 A 
INNER JOIN Table2 B
 on A.CUSTOMER = B.CUSTOMER
AND A.Invoice_number = B.Invoice_number
and concat_ws(' ', A.Amount, 'invoice number', A.invoice_number)= B.Free_Text
where 
AND B.Invoice_number IN ('631','632')
AND A.CUSTOMER = 'A'
AND B.Time_stamp >= TIMESTAMP('2015-01-01 00:00:00')
AND A.Time_stamp >= TIMESTAMP('2015-01-01 00:00:00')

虽然由于字符串连接，性能可能会缓慢。并且无法使用金额指数。

Answer 3

DISTINCT无法解决此问题。

与FREE_Text

相比，您似乎需要根据Ammount限制行数

下面的连接可能需要根据您使用的数据库引擎进行调整。

AND B.FREE_TEXT LIKE '%' + A.Ammount + '%'

或

AND B.FREE_TEXT LIKE CONCAT('%', A.Ammount,'%')

警告：如果您正在使用大量行，这可能效果不佳。

SQL如何删除不正确的重复（不正确）值？

3 个答案: