Trait datafusion_physical_expr::aggregate::AggregateExpr

pub trait AggregateExpr:
    Send
    + Sync
    + Debug
    + PartialEq<dyn Any> {
    // Required methods
    fn as_any(&self) -> &(dyn Any + 'static);
    fn field(&self) -> Result<Field, DataFusionError>;
    fn create_accumulator(
        &self,
    ) -> Result<Box<dyn Accumulator>, DataFusionError>;
    fn state_fields(&self) -> Result<Vec<Field>, DataFusionError>;
    fn expressions(&self) -> Vec<Arc<dyn PhysicalExpr>>;

    // Provided methods
    fn order_bys(&self) -> Option<&[PhysicalSortExpr]> { ... }
    fn order_sensitivity(&self) -> AggregateOrderSensitivity { ... }
    fn with_beneficial_ordering(
        self: Arc<Self>,
        _requirement_satisfied: bool,
    ) -> Result<Option<Arc<dyn AggregateExpr>>, DataFusionError> { ... }
    fn name(&self) -> &str { ... }
    fn groups_accumulator_supported(&self) -> bool { ... }
    fn create_groups_accumulator(
        &self,
    ) -> Result<Box<dyn GroupsAccumulator>, DataFusionError> { ... }
    fn reverse_expr(&self) -> Option<Arc<dyn AggregateExpr>> { ... }
    fn create_sliding_accumulator(
        &self,
    ) -> Result<Box<dyn Accumulator>, DataFusionError> { ... }
    fn all_expressions(&self) -> AggregatePhysicalExpressions { ... }
    fn with_new_expressions(
        &self,
        _args: Vec<Arc<dyn PhysicalExpr>>,
        _order_by_exprs: Vec<Arc<dyn PhysicalExpr>>,
    ) -> Option<Arc<dyn AggregateExpr>> { ... }
    fn get_minmax_desc(&self) -> Option<(Field, bool)> { ... }
}

An aggregate expression that:

• knows its resulting field
• knows how to create its accumulator
• knows its accumulator’s state’s field
• knows the expressions from whose its accumulator will receive values

Any implementation of this trait also needs to implement the PartialEq<dyn Any> to allows comparing equality between the trait objects.

Required Methods

fn as_any(&self) -> &(dyn Any + 'static)

Returns the aggregate expression as Any so that it can be downcast to a specific implementation.

fn field(&self) -> Result<Field, DataFusionError>

the field of the final result of this aggregation.

fn create_accumulator(&self) -> Result<Box<dyn Accumulator>, DataFusionError>

the accumulator used to accumulate values from the expressions. the accumulator expects the same number of arguments as expressions and must return states with the same description as state_fields

fn state_fields(&self) -> Result<Vec<Field>, DataFusionError>

the fields that encapsulate the Accumulator’s state the number of fields here equals the number of states that the accumulator contains

fn expressions(&self) -> Vec<Arc<dyn PhysicalExpr>>

expressions that are passed to the Accumulator. Single-column aggregations such as sum return a single value, others (e.g. cov) return many.

Provided Methods

fn order_bys(&self) -> Option<&[PhysicalSortExpr]>

Order by requirements for the aggregate function By default it is None (there is no requirement) Order-sensitive aggregators, such as FIRST_VALUE(x ORDER BY y) should implement this

fn order_sensitivity(&self) -> AggregateOrderSensitivity

Indicates whether aggregator can produce the correct result with any arbitrary input ordering. By default, we assume that aggregate expressions are order insensitive.

fn with_beneficial_ordering( self: Arc<Self>, _requirement_satisfied: bool, ) -> Result<Option<Arc<dyn AggregateExpr>>, DataFusionError>

Sets the indicator whether ordering requirements of the aggregator is satisfied by its input. If this is not the case, aggregators with order sensitivity AggregateOrderSensitivity::Beneficial can still produce the correct result with possibly more work internally.

Returns

Returns Ok(Some(updated_expr)) if the process completes successfully. If the expression can benefit from existing input ordering, but does not implement the method, returns an error. Order insensitive and hard requirement aggregators return Ok(None).

fn name(&self) -> &str

Human readable name such as "MIN(c2)". The default implementation returns placeholder text.

fn groups_accumulator_supported(&self) -> bool

If the aggregate expression has a specialized GroupsAccumulator implementation. If this returns true, [Self::create_groups_accumulator] will be called.

fn create_groups_accumulator( &self, ) -> Result<Box<dyn GroupsAccumulator>, DataFusionError>

Return a specialized GroupsAccumulator that manages state for all groups.

For maximum performance, a GroupsAccumulator should be implemented in addition to Accumulator.

fn reverse_expr(&self) -> Option<Arc<dyn AggregateExpr>>

Construct an expression that calculates the aggregate in reverse. Typically the “reverse” expression is itself (e.g. SUM, COUNT). For aggregates that do not support calculation in reverse, returns None (which is the default value).

fn create_sliding_accumulator( &self, ) -> Result<Box<dyn Accumulator>, DataFusionError>

Creates accumulator implementation that supports retract

fn all_expressions(&self) -> AggregatePhysicalExpressions

Returns all expressions used in the AggregateExpr. These expressions are (1)function arguments, (2) order by expressions.

fn with_new_expressions( &self, _args: Vec<Arc<dyn PhysicalExpr>>, _order_by_exprs: Vec<Arc<dyn PhysicalExpr>>, ) -> Option<Arc<dyn AggregateExpr>>

Rewrites AggregateExpr, with new expressions given. The argument should be consistent with the return value of the AggregateExpr::all_expressions method. Returns Some(Arc<dyn AggregateExpr>) if re-write is supported, otherwise returns None. TODO: This method only rewrites the PhysicalExprs and does not handle Exprs. This can cause silent bugs and should be fixed in the future (possibly with physical-to-logical conversions).

fn get_minmax_desc(&self) -> Option<(Field, bool)>

If this function is max, return (output_field, true) if the function is min, return (output_field, false) otherwise return None (the default)

output_field is the name of the column produced by this aggregate

Note: this is used to use special aggregate implementations in certain conditions

Implementors

impl AggregateExpr for AggregateFunctionExpr