Skip to main content
cancel
Showing results for 
Search instead for 
Did you mean: 
triciagdaly
Contributor III
Contributor III

Needing quotient of sums of two expressions

I am needing a total for Overall PPV % that is quotient of using sums of  two expressions.  Sum(Ext Variance Amt/PPV) / Sum(Ext Price Paid), but not sure how to accomplish this.    The individual Overall % on the detail lines works fine - Ext Variance Amt/PPV / Ext Price Paid.  In the example attached, using Ext Variance Amt/PPV total of 165,252.01 and Ext Price Paid total of 2,034,568.83, the Overall PPV % should be 7.17%.  Any help would greatly be appreciated.  Also I have a screen shot attached for reference.

3 Replies
hic
Former Employee
Former Employee

You need to think through your formulas a little...

You mix aggregated numbers with unaggregated, and then you will get problems like this.

Image10.png

For instance, in the image above, the [Unburdened Cost] is a reference to another measure, hence it is an aggregated number. But the RECEIPT_PRICE and the quantity are naked (not wrapped in aggregation functions) so they cannot always be evaluated  - especially not on the total row.

You need to wrap these in aggregation functions. E.g. by

=([Unburdened Cost] - Avg(RECEIPT_PRICE))*Sum(fabs(RECEIPT_APPROVED_QUANTITY))

See more on Use Aggregation Functions!

HIC

triciagdaly
Contributor III
Contributor III
Author

Ok.  That makes sense.  The ultimate objective is the quotient of the sum of each of the below divisors.  I will probably need to use aggr, just not very clear on syntax.

([Unburdened Cost] - Avg(RECEIPT_PRICE))*Sum(fabs(RECEIPT_APPROVED_QUANTITY)) / sum(RECEIPT_AMOUNT)

The sum of expression in red / sum of expression in blue is what I’m needing.  Any help is appreciated!

hic
Former Employee
Former Employee

I don't think you need Aggr(). I would go for exactly the formula you've written. However, you will probably need to correct the expression that defines [Unburdened Cost] also (and possibly also the other expressions) since also this contains naked field references. Then I think it will work.

HIC