Executor
Methods
- __construct() : mixed
- execute() : Generator<mixed, PhysicalPlan>
- Plans and runs a whole plan inside one DataFrame span of $context: a planning failure closes the span as failed.
- executePipeline() : Generator<int, Rows>
- Drives one pipeline and the pipelines it reads, all in this process, with no DataFrame span.
- executePlan() : Generator<int, Rows>
- Drives a plan with no DataFrame span - $context belongs to whoever runs it. The config's launcher may launch the head of the plan's chain; without one, or when the launch does not pay, the plan runs here.
- fetch() : Rows
- Executes $logical and merges every batch into one Rows.
- merge() : Rows
- Merges every batch into one Rows. An empty result still carries the schema of the plan the generator returns, or an empty one when that plan cannot describe its rows.
Methods
__construct()
public
__construct() : mixed
execute()
Plans and runs a whole plan inside one DataFrame span of $context: a planning failure closes the span as failed.
public
execute(LogicalPlan $logical, FlowContext $context[, null|RunCounts $counts = null ]) : Generator<mixed, PhysicalPlan>
Parameters
- $logical : LogicalPlan
- $context : FlowContext
- $counts : null|RunCounts = null
-
counts the rows every source yields, a joined frame's included, and the peak of every launched worker
Tags
Return values
Generator<mixed, PhysicalPlan> —returns the physical plan it ran
executePipeline()
Drives one pipeline and the pipelines it reads, all in this process, with no DataFrame span.
public
executePipeline(Pipeline $pipeline, FlowContext $context[, null|Extractor $source = null ][, StepRunner $runner = new StepRunner() ]) : Generator<int, Rows>
Parameters
- $pipeline : Pipeline
- $context : FlowContext
- $source : null|Extractor = null
-
feeds the chain's first pipeline when it reads no source of its own
- $runner : StepRunner = new StepRunner()
-
a branch's own, so the steps its writes become stay the same across its fibers
Tags
Return values
Generator<int, Rows>executePlan()
Drives a plan with no DataFrame span - $context belongs to whoever runs it. The config's launcher may launch the head of the plan's chain; without one, or when the launch does not pay, the plan runs here.
public
executePlan(PhysicalPlan $plan, FlowContext $context[, null|RunCounts $counts = null ]) : Generator<int, Rows>
Parameters
- $plan : PhysicalPlan
- $context : FlowContext
- $counts : null|RunCounts = null
-
counts the rows the plan's source yields and the peak of every launched worker
Return values
Generator<int, Rows>fetch()
Executes $logical and merges every batch into one Rows.
public
fetch(LogicalPlan $logical, FlowContext $context) : Rows
Parameters
- $logical : LogicalPlan
- $context : FlowContext
Tags
Return values
Rowsmerge()
Merges every batch into one Rows. An empty result still carries the schema of the plan the generator returns, or an empty one when that plan cannot describe its rows.
public
merge(Generator<mixed, PhysicalPlan> $batches, Backend $backend) : Rows
Parameters
- $batches : Generator<mixed, PhysicalPlan>
-
what execute() yields
- $backend : Backend