This is exactly what I was trying but having issues with......

It will run on a weekly basis. I have just converted a windows app from nonSQL to SQL. Currently this is done inside the application code and I...

So for now, my solution is to run 3 UPDATE scripts. Each will UPDATE a different NUMERATOR field as follows. The check is in the WHERE statement.

UPDATE...

Works PERFECT!!! Only problem now is that this does not handle when there are more than one records found in my PQRI_main table and thus multiple a.CPT's. When...

Does not work. I am attempting this just on NUMERATOR and all stays NULL. I tried 2 ways:

1:

UPDATE invoices set numerator = CASE NUMERATOR WHEN NULL THEN a.cpt...

Ok, so I did a SELECT using:

SELECT distinct INVOICE, MEASURE

FROM IDXRaw

am get the right number of records. Doesn't that prove that my duplicates do not have extra characters, spaces,...

Yes, I ran the following and was returned with many rows with counts as high as 9.

SELECT INVOICE, MEASURE, COUNT(*)

FROM IDXRaw

GROUP BY INVOICE, MEASURE

IDXRaw is my source table.

Grasshopper,

I am pretty darn definite that there are duplicates based on the 2 fields I am considering (DIVISION and INVNUM.) There are something like 1.5 million records I am...

I am guessing it is not seeing the duplicates from this mass INSERT because they are not yet committed? If so, is there a work around to that?

except there are duplicates from the 1st run. The SOURCE table has duplicates. So when record 1 is inserted into TARGET table and then record 3 which is...

So I must not understand how SQL SERVER does inserts in this kind of script. Whe table is empty and I run the previous script to insert 700,000 records...

Pablo,

Thanks for the response but this does not seem to work. I am doing the following and still getting duplicated recoreds with same INVNUM and MEASURE.

INSERT INTO invoices (INVNUM,...