one sort of brute force method to do it would be to sort first by invoice
(column a?)and secondly by error (column C?) and assuming your data starts in
row 2. add a helper column (column D?) and enter in D2
=if(and(A1=A2,C1=C2),D1+1,1)
and copy down to the bottom of the data
In E2 enter
=max(offset(d2,0,0,countif(A:A,A2))
in E3
=if(A2=A3,E2,max(offset(d3,0,0,countif(A:A,A3)))
and copy to the bottom of the data
In F2
=if(D2=E2,1,'')
and copy to the bottom of the data
use auto filter to first select only ones in column F and the invoice of
interest in coulme A this will give you the errors which occur the most in
each invoice. Note if two or three errors happen the same number of times,
they will all show up.

"Jamesy" wrote:

I find myself really struggling with this one and in the past you guys have
all ways come up with the goods, so don't let me down this time!

Let me explain my problem in simple form: -

1.) I start with an invoice
2.) One invoice holds many product lines
3.) An invoice fails, on this invoice there could be 1 line, 2 lines or 20
lines failing
4.) Within this invoice (failing product lines) alpha numeric codes are
provided to help us understand how to fix these individual problems

My end goal is to understand an average mode failure for each invoice. Let
me provide an example invoice: -

Invoice number Product Error
123456 Coke can AB123
123456 Lemonade can AC123
123456 Milk carton AC123
123456 Juice container AC123

Now from the above even though there are 4 lines this is actually one
invoice, when you uniquely identify the invoice numbers I'll get the top one
and when I look across I'll see error code AB123 which isn't a fair
reflection for this invoice.

Any help/advice would be warmly welcomed!

Thanks,
James.