Struct ParquetLookupSource 

pub struct ParquetLookupSource { /* private fields */ }

A LookupSource that loads a Parquet file into memory.

All data is loaded eagerly at construction time. Lookups are served from an in-memory HashMap keyed by the serialized primary key columns. The values are single-row RecordBatch instances.

This is appropriate for dimension tables that fit in memory and are read-mostly or static.

Implementations

impl ParquetLookupSource

pub fn from_path(config: ParquetLookupSourceConfig) -> Result<Self, LookupError>

Load a Parquet file from the local filesystem.

Reads all record batches, extracts primary key columns, and stores each row in an in-memory HashMap.

Errors

Returns LookupError::Internal if the file cannot be opened, read, or if the primary key columns are not found in the schema.

pub fn schema(&self) -> &SchemaRef

Returns the Arrow schema of the loaded Parquet file.

Trait Implementations

impl Debug for ParquetLookupSource

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

Formats the value using the given formatter.

impl LookupSource for ParquetLookupSource

fn query( &self, keys: &[&[u8]], _predicates: &[Predicate], _projection: &[ColumnId], ) -> impl Future<Output = Result<Vec<Option<RecordBatch>>, LookupError>> + Send

Query the source by keys, predicates, and/or projection.

fn capabilities(&self) -> LookupSourceCapabilities

Capabilities this source advertises.

fn source_name(&self) -> &str

Source name for logging and metrics.

fn schema(&self) -> SchemaRef

Arrow schema of the data this source returns.

fn estimated_row_count(&self) -> Option<u64>

Optional row count estimate for query planning.

async fn health_check(&self) -> Result<(), LookupError>

Health check. Default: always healthy.