Skip to content

Package 7 — Choice Components

Package 7 defines DataChoice, a disjoint union type for modeling alternatives in a data stream.

DataChoice

A DataChoice describes a set of named component alternatives — at runtime, exactly one of the items is present. The choiceValue field acts as a discriminator in the encoded data stream, telling the decoder which alternative was selected.

Field Type Description
dataComponent AbstractDataComponent Inherited metadata
choiceValue Category Discriminator — its value names the selected item
items list of NamedComponent The available alternatives (min 1)

How it works

The choiceValue is a Category component whose valid tokens correspond to the name fields of the items. When encoding, the value of choiceValue tells the reader which item structure follows in the stream.

Example: multi-format sensor output

A sensor that can output either a scalar reading or a full spectrum:

DataChoice "SensorOutput"
  choiceValue: Category (values: ["scalar", "spectrum"])
  items:
  ├── scalar: Quantity (uom: "W/m2")
  └── spectrum: DataArray
        elementType: Quantity (uom: "W/m2/nm")
        elementCount: 256

DataChoice vs. language-level unions

The AnyComponent union (Package 5) is a schema-level construct — it's built into the IDL and resolved at deserialization time. DataChoice is a data-level construct — it lives inside the data stream and is resolved by reading the discriminator value at runtime. Use DataChoice when the set of alternatives is defined by the data model, not by the schema.

struct DataChoice {
  dataComponent @0 :BT.AbstractDataComponent;
  choiceValue   @1 :SC.Category;
  items         @2 :List(NamedComponent);
}

table DataChoice {
  data_component:AbstractDataComponent;
  choice_value:Category;
  items:[NamedComponent] (required);
}

message DataChoice {
  AbstractDataComponent data_component = 1;
  Category choice_value = 2;
  repeated NamedComponent items = 3;
}