Expand description
Build a differential-dataflow graph from compiled PlanNodes and run a
whole Model to a set of computed derived-measure values.
This connects the formula compiler (compiler) to the DD substrate proven
by the spike in lib.rs. Encoding is the same the spike validated:
- DD key = serialized coordinate
CoordKey=Vec<(u32,u32)> - DD value =
f64::to_bits()asu64(numeric measures only, v1) - diff
R=isize; a finalreducecollapses to one value per key.
Scope (v1, numeric): InputMeasure, Literal, MapUnary(Neg), MapBinary
(arithmetic + comparison/logical encoded as 1.0/0.0), Join (dimension
alignment), Aggregate (SUM/AVG/MIN/MAX), and scalar FuncCall built-ins
(ABS/ROUND/FLOOR/CEIL/SQRT/NEG, MIN2/MAX2). Non-numeric derived values
(Text/Boolean stored as such) and Date values are carried in CellValue;
comparison/logical results live in the numeric lane as 1.0/0.0.
Enums§
Functions§
- evaluate
- Run
modelto completion and return computed values for the requested derived measures:measure -> (coordinate-key -> value).