Crate vortex_datafusion 

Integrations between Vortex and DataFusion.

The crate exposes two main entry points:

• VortexFormatFactory for the file-based integration used by SQL, CREATE EXTERNAL TABLE, and ListingTable.
• v2 for direct integration from an existing Vortex DataSourceRef.

Registering The File Format

Most applications register VortexFormatFactory with a DataFusion SessionContext and then let DataFusion create VortexFormat and VortexSource instances as queries are planned:

use std::sync::Arc;

use datafusion::datasource::provider::DefaultTableFactory;
use datafusion::execution::SessionStateBuilder;
use datafusion::prelude::SessionContext;
use datafusion_common::GetExt;
use vortex_datafusion::VortexFormatFactory;

let factory = Arc::new(VortexFormatFactory::new());
let mut state_builder = SessionStateBuilder::new()
    .with_default_features()
    .with_table_factory(
        factory.get_ext().to_uppercase(),
        Arc::new(DefaultTableFactory::new()),
    );

if let Some(file_formats) = state_builder.file_formats() {
    file_formats.push(factory.clone() as _);
}

let ctx = SessionContext::new_with_state(state_builder.build()).enable_url_table();
ctx.sql(
    "CREATE EXTERNAL TABLE metrics (service VARCHAR, value BIGINT) \
     STORED AS vortex LOCATION 'file:///tmp/metrics/'",
)
.await?;

Registering An Existing Vortex Data Source

If your application already has a Vortex DataSourceRef, use v2::VortexTable to register it directly with DataFusion:

use std::sync::Arc;

use arrow_schema::Schema;
use datafusion::prelude::SessionContext;
use vortex::VortexSessionDefault;
use vortex::scan::DataSourceRef;
use vortex::session::VortexSession;
use vortex_datafusion::v2::VortexTable;

let table = Arc::new(VortexTable::new(
    data_source,
    VortexSession::default(),
    Arc::new(Schema::empty()),
));

let ctx = SessionContext::new();
ctx.register_table("vortex_data", table)?;

Modules

metrics

Helpers for extracting Vortex scan metrics from DataFusion execution plans.

reader

Factory for creating VortexReadAt instances from PartitionedFiles.

v2

Direct DataFusion integration for an existing Vortex DataSourceRef.

Structs

VortexAccessPlan

Additional Vortex-specific scan constraints attached to a PartitionedFile.

VortexFormat

DataFusion FileFormat implementation for .vortex files.

VortexFormatFactory

Registration entry point for the file-backed Vortex integration.

VortexSource

File scan implementation for reading one or more .vortex files.

VortexTableOptions

Options to configure VortexFormat and VortexSource.

Traits

ExpressionConvertor

Trait for converting DataFusion expressions to Vortex ones.