Struct ClickHouseBuilder 

pub struct ClickHouseBuilder { /* private fields */ }

Entrypoint builder for ClickHouse and DataFusion integration.

Ensures DataFusion default UDFs as well as ClickHouse udfs are registered in SessionContext, the pool is attached to the ClickHouseTableProviderFactory, and ClickHouse is reachable by the provided endpoint.

NOTE: While clickhouse_arrow defaults to binary encoding for strings (via clickhouse_arrow::ArrowOptions::strings_as_strings == false), for DataFusion the default is true. This can be disabled by modifying the setting via ClickHouseBuilder::configure_arrow_options

Implementations

impl ClickHouseBuilder

pub fn new(endpoint: impl Into<Destination>) -> Self

Create new ClickHouseBuilder that can be used to configure both the clickhouse_arrow::ClientBuilder, the clickhouse_arrow::ArrowPoolBuilder, and be “built” into a ClickHouseCatalogBuilder for interacting with ClickHouse databases and tables.

pub fn new_with_pool_builder( endpoint: impl Into<Destination>, pool_builder: ArrowConnectionPoolBuilder, ) -> Self

pub fn with_coercion(self, coerce: bool) -> Self

Configure whether to coerce schemas to match expected schemas during query execution. Remember, there is a non-zero cost incurred per RecordBatch, and this is mainly useful for allowing looser return types when using clickhouse functions.

pub fn configure_client( self, f: impl FnOnce(ClientBuilder) -> ClientBuilder, ) -> Self

Configure the underlying clickhouse_arrow::ClientBuilder.

pub fn configure_pool( self, f: impl FnOnce(ArrowPoolBuilder) -> ArrowPoolBuilder, ) -> Self

Configure the underlying clickhouse_arrow::ArrowPoolBuilder.

pub fn configure_arrow_options( self, f: impl FnOnce(ArrowOptions) -> ArrowOptions, ) -> Self

clickhouse_arrow defaults to binary encoding for datafusion::arrow::datatypes::DataType::Utf8 columns, the default is inverted for DataFusion. This method allows disabling that change to use binary encoding for strings, among other configuration options.

See: clickhouse_arrow::ArrowOptions

pub async fn build_catalog( self, ctx: &SessionContext, catalog: Option<&str>, ) -> Result<ClickHouseCatalogBuilder>

Build ensures ClickHouse builtins are registered (such as nested functions), the pool is attached to the factory, the ClickHouse endpoint is reachable, and the catalog is created and registered to the SessionContext.

Errors

• Returns an error if the ClickHouse endpoint is unreachable
• Returns an error if the ClickHouse catalog fails to be created