Challenge 2: A Safe-Division DAX Measure — Solution Walkthrough

DAX measure:
  Average Order Value :=
  DIVIDE(SUM(Sales[Amount]), SUM(Sales[OrderCount]))

Walkthrough:

SUM(Sales[Amount]) totals every order's amount for whatever filter
context is currently active (e.g. a specific Region, if the
PivotTable groups by Region). SUM(Sales[OrderCount]) totals the
number of orders the same way. DIVIDE wraps the two, dividing the
first by the second, and — because DIVIDE is DAX's own safe-division
function — returns blank instead of an error if the order count
happens to be zero for some grouping (e.g. a Region with no orders
at all in the selected period), rather than showing #DIV/0!.

What would go wrong trying to build this as a Chapter 3-style calculated field:

Nothing would technically prevent TYPING a calculated field with the
same division — Sales Amount / Sales OrderCount — but two real
problems would follow: first, a calculated field has no equivalent of
DIVIDE, so a zero-order group would show a raw #DIV/0! error instead
of a clean blank, requiring an awkward workaround. Second, and more
fundamentally, a calculated field can only be defined and used within
ONE specific PivotTable that has both Sales fields directly available
to it — it cannot be defined once in the Data Model and reused
automatically by every other PivotTable built from the same model the
way a DAX measure can. If Category (from Products) needs to appear
alongside this average in a different PivotTable, a calculated field
would have no way to reach across to a related table at all, whereas
the DAX measure works immediately in any PivotTable built from the
Data Model, regardless of which related table's fields sit alongside
it.

Notes:
- This exercise is a direct, concrete illustration of this chapter's
  own comparison table: "works across related tables" and "built-in
  safe division" are exactly the two rows where DAX measures out-
  perform calculated fields, and this specific formula needs both.