Hi Claus,

I have growing frustration over what appears to be a moving target of what the final out come is supposed to be, and of course I'm in over my head on the code.

Raw data in column A and as before move the non-P item up and over 1 row and 1 column.

So now data is in column A and column B (which is not shown below.)

Count the number of identical P-xxxx's with nothing in column B next to them and the number of identical P-xxxx's with ABCxxxx in column B next to them.

P-1234 ABC and P-1234 ABC would be 2.
P-4567 DEF and P-4567 HIJ would be 1 for each.
P-1357 XXX and P-1388 XXX would be 1 for each.

Those sum go in column C.

P-4352
P-3534
P-4568
ABCDE123
P-3333
P-5506
CDEFG234
P-4352
P-3534
P-4568
ABCDE124
P-7679
P-9852


P-9876 ABCDE 1
P-5678 FGRTTTRGF 1
P-7675 HFHFHFH 1
P-8901 ABCDE 1
P-8901 4
P-3456 1
P-6543 2
P-54463 1

The item in column B is a "serial number" so most likely all the P-xxxx with a
serial number in column C will be 1. But if you have a couple or three
P-9876 ABCDE's, for example, then that number should be in column C.

Notice P-8901 ABCDE and P-8901 are counted separate.

Once the count is completed then the duplicates can be removes to show as the column A, B, C example.

The data above shows format only. None of the final outcome data is actually part of the raw data. It is just a before and after view so to speak.

This is my last shot at this caper, I feel I've abused the privilege of the news group and your patience too much.

Thanks.

Howard