many times when constructing a data model you need to
make a join between Header and Details tables into one new table(i know you can avoid it in some cases but let's assume you cant)
for example:
invoice_header joined with invoice_details (on customer and invoice number) to make a new INVOICES table.
this has the side effect of multiplying the values in each row of the header times the number of corresponding lines in the details.
so if you need to perform some aggregation on a field originating from the header you must deal with this multiplication.
usually by using one of these expressions: (simplified for simplicity 🙂 )
sum(aggr(avg(Invoice_amount),Customer))
OR
sum(aggr(max(Invoice_amount),Customer,invoice))
the question I'm trying to raise for discussion is which is faster?
as I see it,
when using the AVG version qlik need to sum the amount and divide by the number of lines
but when using the max version qlik needs to sort the values to find the max
is it possible there is no difference?
do you think any is faster, if so why?
or you think this is a redundant unimportant question and you have better things to spend your time on 🙂
Daniel
