Module mz_compute_types::plan

An explicit representation of a rendering plan for provided dataflows.

Modules

• flat_plan
  FlatPlan, a flat representation of LIR plans used in the compute protocol and rendering, and support for for flatting Plans into this representation.
• interpret
  Utilities and implementations for abstract interpretation of crate::plan::Plan structures.
• join
  Planning of Plan::Join operators, and supporting types.
• lowering 🔒
  Lowering DataflowDescriptions from MIR (MirRelationExpr) to LIR (Plan).
• proto_available_collections
  Nested message and enum types in ProtoAvailableCollections.
• proto_get_plan
  Nested message and enum types in ProtoGetPlan.
• reduce
  Reduction execution planning and dataflow construction. We build ReducePlans to manage the complexity of planning the generated dataflow for a given reduce expression. The intent here is that each creating a ReducePlan should capture all of the decision making about what kind of dataflow do we need to render and what each operator needs to do, and then actually rendering the plan can be a relatively simple application of (as much as possible) straight line code.
• threshold
  Threshold planning logic.
• top_k
  TopK planning logic.
• transform
  Utilities and implementations for transformation of crate::plan::Plan structures.

Structs

• AvailableCollections
  The forms in which an operator’s output is available; it can be considered the plan-time equivalent of render::context::CollectionBundle.
• ProtoAvailableCollections
• ProtoGetPlan
• ProtoLetRecLimit

Enums

• GetPlan
  How a Get stage will be rendered.
• Plan
  A rendering plan with as much conditional logic as possible removed.

Constants

• _IMPL_ARBITRARY_FOR_AvailableCollections 🔒
• _IMPL_ARBITRARY_FOR_GetPlan 🔒

Functions

• any_arranged_thin 🔒
  A strategy that produces arrangements that are thinner than the default. That is the number of direct children is limited to a maximum of 3.
• bucketing_of_expected_group_size 🔒
  Returns bucket sizes, descending, suitable for hierarchical decomposition of an operator, based on the expected number of rows that will have the same group key.

Type Aliases

• LirId
  An identifier for an LIR node.