Struct DatafusionArrowPredicate 

pub(crate) struct DatafusionArrowPredicate {
    physical_expr: Arc<dyn PhysicalExpr>,
    projection_mask: ProjectionMask,
    rows_pruned: Count,
    rows_matched: Count,
    time: Time,
    schema_mapper: Arc<dyn SchemaMapper>,
}

A “compiled” predicate passed to ParquetRecordBatchStream to perform row-level filtering during parquet decoding.

See the module level documentation for more information.

Implements the ArrowPredicate trait used by the parquet decoder

An expression can be evaluated as a DatafusionArrowPredicate if it:

• Does not reference any projected columns
• Does not reference columns with non-primitive types (e.g. structs / lists)

Fields

physical_expr: Arc<dyn PhysicalExpr>

the filter expression

projection_mask: ProjectionMask

Path to the columns in the parquet schema required to evaluate the expression

rows_pruned: Count

how many rows were filtered out by this predicate

rows_matched: Count

how many rows passed this predicate

time: Time

how long was spent evaluating this predicate

schema_mapper: Arc<dyn SchemaMapper>

used to perform type coercion while filtering rows

Implementations

impl DatafusionArrowPredicate

pub fn try_new( candidate: FilterCandidate, metadata: &ParquetMetaData, rows_pruned: Count, rows_matched: Count, time: Time, ) -> Result<Self>

Create a new DatafusionArrowPredicate from a FilterCandidate

Trait Implementations

impl ArrowPredicate for DatafusionArrowPredicate

fn projection(&self) -> &ProjectionMask

Returns the [ProjectionMask] that describes the columns required to evaluate this predicate.

fn evaluate(&mut self, batch: RecordBatch) -> ArrowResult<BooleanArray>

Evaluate this predicate for the given [RecordBatch] containing the columns identified by [Self::projection]

impl Debug for DatafusionArrowPredicate

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

Formats the value using the given formatter.