Struct FilePruner 

pub struct FilePruner { /* private fields */ }

Prune based on file-level statistics.

Note: Partition column pruning is handled earlier via replace_columns_with_literals which substitutes partition column references with their literal values before the predicate reaches this pruner.

Implementations

impl FilePruner

pub fn new( predicate: Arc<dyn PhysicalExpr>, logical_file_schema: &SchemaRef, _partition_fields: Vec<FieldRef>, partitioned_file: PartitionedFile, predicate_creation_errors: Count, ) -> Result<Self>

👎Deprecated since 52.0.0:

Use try_new instead which returns None if no statistics are available

pub fn try_new( predicate: Arc<dyn PhysicalExpr>, file_schema: &SchemaRef, partitioned_file: &PartitionedFile, predicate_creation_errors: Count, ) -> Option<Self>

Create a file pruner for this file, or None when pruning it cannot help.

Returns None when the file has no statistics struct to evaluate a pruning predicate against, or when the predicate is purely static and the file has no usable column statistics — in that case planning already did everything such a pruner could. A predicate carrying a dynamic filter is always accepted (given a statistics struct), since it may prune via partition-value folding even without column statistics.

pub fn is_watching(&self) -> bool

Returns true if this pruner watches a dynamic filter that can still change, meaning Self::should_prune is worth re-checking as the scan progresses. When false, the predicate is effectively static for the remainder of the scan and the caller can avoid wrapping the stream in a per-batch re-pruning adapter.

pub fn should_prune(&mut self) -> Result<bool>