Add documentation for distinct count optimizations in DAX

Author: marcosqlbi

…s/optimization-notes/dimensions-joins.md …ls/optimization-notes/dimension-joins.md_mydocs/dax-internals/optimization-notes/dimensions-joins.md renamed to _mydocs/dax-internals/optimization-notes/dimension-joins.md _mydocs/dax-internals/optimization-notes/dimensions-joins.md renamed to _mydocs/dax-internals/optimization-notes/dimension-joins.md

@@ -0,0 +1,23 @@
+---
+layout:     page
+title:      Distinct Count
+published:  true
+order:      /60
+---
+
+The DISTINCTCOUNT function in DAX can generate different storage engine queries depending on the complete DAX query. This page includes special notes about storage engine optimization.
+
+## DISTINCTCOUNT on primary keys
+
+The DISTINCTCOUNT function in DAX is automatically converted to a COUNTROWS when the argument is a key column of the table.
+For example, consider the following expression: 
+```DAX
+DISTINCTCOUNT ( 'Product'[ProductKey] )
+```
+
+If *ProductKey* is on the one-side of a regular one-to-many relationship, or it is marked as a key column of the table, then the code corresponds to: 
+```DAX
+COUNTROWS ( 'Product' )
+```
+
+Because COUNTROWS usually produces a more efficient query plan, this is usually an optimization that does not have side effects.