Enum NodeOp 

pub enum NodeOp {
    Scan {
        table: String,
        filters: Vec<String>,
    },
    Filter {
        predicate: String,
    },
    Project {
        columns: Vec<String>,
    },
    Aggregate {
        group_keys: Vec<String>,
    },
    Join {
        join_type: JoinType,
    },
    Exchange {
        partitioning: Partitioning,
    },
    Sink {
        format: String,
    },
    CoalescePartitions {
        target_partitions: usize,
    },
    CreateLiveTable {
        name: String,
        query: String,
    },
    RefreshLiveTable {
        name: String,
    },
    DropLiveTable {
        name: String,
    },
    KeyBy {
        key_column: String,
    },
    Watermark {
        event_time_column: String,
        lag_ms: u64,
    },
    Window {
        spec: Box<WindowExecutionSpec>,
    },
    StreamSource {
        source_id: String,
        bounded: bool,
    },
    StateTtl {
        ttl_ms: u64,
    },
    GlobalSort {
        keys: Vec<(String, bool)>,
    },
    SortMergeJoin {
        join_type: JoinType,
        left_keys: Vec<String>,
        right_keys: Vec<String>,
    },
    WindowJoin {
        join_type: JoinType,
        left_keys: Vec<String>,
        right_keys: Vec<String>,
        time_column: String,
        window_ms: u64,
    },
    Unnest {
        array_column: String,
        output_column: String,
        with_ordinality: bool,
    },
    Cep {
        key_column: String,
        event_time_column: String,
        stage_column: String,
    },
    SkewJoin {
        keys: Vec<String>,
        factor: u32,
        join_type: JoinType,
    },
    Other {
        description: String,
    },
}

Typed operator classification for a plan node.

Variants

Scan

Table or file scan, with optional pushed-down filter predicates.

Fields

table: String

filters: Vec<String>

Filter

Row filter with a predicate expression string.

Fields

predicate: String

Project

Column projection.

Fields

columns: Vec<String>

Aggregate

Aggregation with optional group keys.

Fields

group_keys: Vec<String>

Join

Join of two inputs.

Fields

join_type: JoinType

Exchange

Data exchange / shuffle between partitions.

Fields

partitioning: Partitioning

Sink

Output sink.

Fields

format: String

CoalescePartitions

AQE coalesce: merge many small partitions into fewer larger ones.

Inserted by the AQE CoalesceRule when runtime statistics show that partition count can be reduced to improve downstream task efficiency.

Fields

target_partitions: usize

Number of output partitions after coalescing.

CreateLiveTable

Create a live table backed by a streaming query.

Fields

name: String

query: String

RefreshLiveTable

Refresh materialized state for a live table.

Fields

name: String

DropLiveTable

Drop a live table.

Fields

name: String

KeyBy

Key stream by column before windowing.

Fields

key_column: String

Watermark

Event-time watermark on a keyed stream.

Fields

event_time_column: String

lag_ms: u64

Window

Windowed streaming operator (tumbling, sliding, or session window).

Fields

spec: Box<WindowExecutionSpec>

StreamSource

Bounded or unbounded stream source.

Fields

source_id: String

bounded: bool

StateTtl

Operator state TTL for streaming nodes.

Fields

ttl_ms: u64

GlobalSort

E2.2: Globally-sorted output produced by a three-stage pipeline: local sort → range-partition shuffle → merge-sort. The executor treats this as a batch pipeline that produces a single sorted partition.

Fields

keys: Vec<(String, bool)>

Ordered list of (column, ascending) sort keys.

SortMergeJoin

E2.2 / E2.4: Sort-merge join using pre-sorted, range-partitioned inputs.

Fields

join_type: JoinType

left_keys: Vec<String>

Column names used as equi-join keys (must match sort order).

right_keys: Vec<String>

WindowJoin

E3.2: Time-windowed join: buffer both streams in the window interval, emit matched pairs when the window closes.

Fields

join_type: JoinType

left_keys: Vec<String>

Column names used as equi-join keys.

right_keys: Vec<String>

time_column: String

Event-time column used to determine window membership.

window_ms: u64

Window duration in milliseconds.

Unnest

E5.2: Expand an array-typed column into one row per element.

Equivalent to UNNEST(array_column) in SQL or a LATERAL join over an array. The output_column name is used for the expanded element. If with_ordinality is true an extra ordinality column (u64) is appended with the 1-based position of each element.

Fields

array_column: String

output_column: String

with_ordinality: bool

Cep

CEP sequential pattern match on a keyed stream.

stage_column names the column whose string value identifies which pattern stage each row belongs to. The executor groups rows by key_column, routes each row to PartitionedCepMatcher::process_event with the row’s stage name, and emits concatenated match batches.

Fields

key_column: String

event_time_column: String

stage_column: String

SkewJoin

AQE skew mitigation: split a hot partition into N sub-partitions by appending a salt column to the join key. The build side is replicated factor times so that each salted sub-partition of the probe side joins against the full build side in parallel. The unsalt node strips the salt column from the post-join output.

Equivalent to Spark AQE’s OptimizeSkewedJoin rule.

Fields

keys: Vec<String>

The join key columns on both sides (must match).

factor: u32

Number of sub-partitions the hot side is split into.

join_type: JoinType

Original join type — kept so the executor can dispatch correctly.

Other

Operator not covered by the above variants.

Fields

description: String

Trait Implementations

impl Clone for NodeOp

fn clone(&self) -> NodeOp

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for NodeOp

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for NodeOp

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Eq for NodeOp

impl PartialEq for NodeOp

fn eq(&self, other: &NodeOp) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl Serialize for NodeOp

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for NodeOp