>>>>Hi,
>>>>
>>>>I have a table where I want to delete ONE of two duplicate records based on the fact that values in two fields are the same. For example:
>>>>
>>>>

>>>>FIELD1    FIELD2
>>>>1              Entry ABC
>>>>1              Entry ABC
>>>>2              Entry 123
>>>>2              Entry 123
>>>>

>>>>
>>>>and so on. There are other fields in this table but they should not be considered when deleting one record. How can you do it in T-SQL?
>>>>
>>>>TIA.
>>>
>>>

>>>DECLARE @TEst TABLE (FIELD1 int, FIELD2 char(10))
>>>INSERT INTO @Test VALUES (1,'Entry ABC')
>>>INSERT INTO @Test VALUES (1,'Entry ABC')
>>>INSERT INTO @Test VALUES (2,'Entry 123')
>>>INSERT INTO @Test VALUES (2,'Entry 123')
>>>
>>>;with cteTest
>>>AS
>>>(
>>>SELECT *, ROW_NUMBER() OVER (PARTITION BY Field1, Field1 ORDER BY Field1, Field1) AS R_C
>>>FROM @TEst Test
>>>)
>>>
>>>DELETE FROM cteTest WHERE R_C > 1
>>>
>>>SELECT * FROM @TEst 
>>>

>>
>>Since I have to do this on a customer's table and I don't understand the syntax, if you don't mind, let me ask you a follow up question.
>>What specific thing - in the customer table - would prevent the above approach to work correctly?
>
>You should list ALL fields that define records as duplicated in
>

>(PARTITION BY Field1, Field2 ORDER BY Field1, Field2)
>

>So if you have 3 or more they all should be here:
>

>(PARTITION BY Field1, Field2, Field3...)
>

>And you should define what records must be deleted
>that is the ORDER BY clause
>

>(PARTITION BY Field1, Field2 ORDER BY SomeOtherFIeldIfYouWant)
>

>
>
>This example define duplicate records by Field1 and FIeld2 and leave
>earliest record defined by datetime field:
>

>DECLARE @TEst TABLE (FIELD1 int, FIELD2 char(10), Field3 DATETIME)
>INSERT INTO @Test VALUES (1,'Entry ABC', GETDATE())
>waitfor delay '00:00:01'
>INSERT INTO @Test VALUES (1,'Entry ABC', GETDATE())
>waitfor delay '00:00:01'
>INSERT INTO @Test VALUES (2,'Entry 123', GETDATE())
>waitfor delay '00:00:01'
>INSERT INTO @Test VALUES (2,'Entry 123', GETDATE())
>SELECT * FROM @TEst 
>
>;with cteTest
>AS
>(
>SELECT *, ROW_NUMBER() OVER (PARTITION BY Field1, Field1 ORDER BY Field3) AS R_C
>FROM @TEst Test
>)
>
>DELETE FROM cteTest WHERE R_C > 1
>
>SELECT * FROM @TEst 
>

This is pretty straight forward. Thank you for the explanation. Of course I will back up the customer table first, before deleting.