Learning: Nodes

• Projection Nodes are used to:
  select columns, define calculated columns, parameters that request values at runtime, and filter on the data source.

• Aggregation Node:
  Used to aggregate functions to measures based on one or several attributes.
  In an Aggregation Node, a calculated column is always computed after the aggregate function.
  Keep Flag: allows to trigger the calculation at the relevant level of granularity.
  Transparent Filter: used when stacked view where the lower views have distinct count measures or upper calculation views contain filters on columns that are not projected.

• Join Node:
  Joins vs Aggregation :: Regular Join, the aggregation is executed after the join. Dynamic Join, aggregation triggered to remove the column if the column is not requested by the client query.
  
  Multi Join order: Outside-In (the table that is furthest from the central table is executed first) or Inside-Out (the table that is closest to the central table is executed first).