Set analysis inheritance

[#1]

Hey all,

I'm writing a couple of variable to calculate KPI's and min/avg/max values of thoose KPI's. The min/avg/max should be calculated ignoring certain selections in the qvw. The problem is that the written functions can get very long. For example:

KPI = count(value)*(max(date)-min(date))

When I need to remove the filters to calculate min/avg/max then I need to use something like:

KPI_avg = avg({$<removeValue=>}count({$<removeValue=>}value)*(max({$<removeValue=>}date)-min({$<removeValue=>}date)))

I have a longer function for my KPI and an even longer selection I need to ignore.

So my question is: can I let the set analysis be inherited by defining my 'ignore selection' only once? for ex:

KPI_avg_wanted = avg({$<removeValue=>}count(value)*(max(date)-min(date)))

I tried something similar to the above but that doesn't work. Anyone got suggestions? thanks!

[jagan]

Hi,

You need to give set analysis expression for every individual method you use in the expression.  So, I don't think it is possible, you have to specify like this

KPI_avg = avg({$<removeValue=>}count({$<removeValue=>}value)*(max({$<removeValue=>}date)-min({$<removeValue=>}date)))

Regards,

Jagan.

[#3]

Thanks, that's what I thought but I had hoped to be wrong

[swuehl]

I believe something like

= avg({$<removeValue=>}count({$<removeValue=>}value)* max({$<removeValue=>}date)-min({$<removeValue=>}date)))

shouldn't work at all, since you are embedding aggregation functions into another aggregation function. You need to use advanced aggregation to handle this.