As a simple starter approach, you can create a Straight Table with a calculated Dimension of "Vendor & Amount" and an expression like "count(InvoiceNumber)". Those with 2 or more are potential duplicates. You can further refine this with Set Analysis to show only those where Count(InvoiceNumber)>1.
You could do something similar in script. It's hard to give specific script without seeing your data model. Can you post a screenshot of your table viewer?
I appreciate both your quick replies and will check into what you said. Rob, below is my data model and the fields I mentioned are highlighted - what's tricky about the invoice number is there will never be an exact duplicate since the system doesn't let invoice numbers be duplicated, but users get around that by changing one character or appending a character on the end. I don't think that Qlikview has any type of fuzzy searching capability like that, which is why I'm leaving that piece out.
If you want to search the invoices with exact same amount and same Vendor, I think Rob's solution should work perfectly fine to identify the candidates.
You would then need to inspect the invoice numbers or other attributes to check if they are real duplicates.
I have somehow misread your request in a way that amounts could also slightly differ, hence I suggested above more graphical approach. To get the invoice number on the scatter graph axis, it needs to be a number (or you may use another appropriate attribute, like the invoice timestamp). As said, I think Rob's solution is better if you are searching same amounts.
If you are interested in string distance functions to compare the invoice numbers, you can start with
AFAIR, there are some threads coping with the Levenshtein algorithm,
If users only append characters, there might be easier solutions (match the smaller string with the beginning of larger string).