ActivePivot context dependend formatting

Question

I have a price measure that I want to format differently in different locations, e.g. the formatting depends on the value of a dimension. I could have multiple measures with different formatting, but that would complicate a lot of other things.

The standard IFormatter framework does not have location information as far as I understand.

I guess could implement this as a post processed mesaure? I could maybe also have different object types for the different types of prices, and then use an IFormatter implementation that switched on the object type?

Is there a best practice for this - pros/cons?

Answer 1

In ActivePivot a formatter does not know the cube location for which a cell has been generated (actually with MDX constructs some cell values can be computed that are not associated to any location).

But as ActivePivot is object based there is always a solution to do what you want:

For the price measure you could write a "Price" java class that contains an amount and a currency, instead of just using a number. Out of the box ActivePivot will not know how to aggregate those prices, so you'll also write a custom aggregation function (extending com.quartetfs.biz.pivot.aggfun.impl.GenericAggregationFunction).

Then you can implement the toString() method of your Price class, as ActivePivot uses this for default formatting. And you'll have access to the currency when formatting. Of course if you want a more decoupled design, or user-based contextual formatting you can also write a PriceFormatter to format your price instances.