Query

geneva.query.GenevaQuery

Bases: BaseModel

base

base: Query

shuffle

shuffle: bool | None = None

shuffle_seed

shuffle_seed: int | None = None

fragment_ids

fragment_ids: list[int] | None = None

with_row_address

with_row_address: bool | None = None

column_udfs

column_udfs: list[ColumnUDF] | None = None

source_table

source_table: str | None = None

source_db_uri

source_db_uri: str | None = None

extract_column_udfs

extract_column_udfs(
    packager: UDFPackager,
) -> list[ExtractedTransform]

Loads a set of transforms that reflect the column_udfs and map_batches_udfs of the query.

geneva.query.GenevaQueryBuilder

Bases: LanceEmptyQueryBuilder

A proxy that wraps LanceQueryBuilder and adds geneva-specific functionality.

schema

schema: Schema

select

select(
    columns: list[str] | Mapping[str, str | UDF | Expr],
) -> Self

Select the output columns of the query.

Parameters:

• columns (list[str] | Mapping[str, str | UDF | Expr]) –

  The columns to select.

  If a list of strings, each string is the name of a column to select.

  If a dictionary of strings then the key is the output name of the column and the value is either an SQL expression (str), an Expr, or a UDF.

shuffle

shuffle(seed: int | None = None) -> Self

Shuffle the rows of the table

enable_internal_api

enable_internal_api() -> Self

Enable internal APIs WARNING: Internal APIs are subject to change

with_fragments

with_fragments(fragments: list[int] | int) -> Self

Filter the rows of the table to only include the specified fragments.

with_row_address

with_row_address() -> Self

Include the physical row address in the result WARNING: INTERNAL API DETAIL

with_read_version

with_read_version(version: int | None) -> Self

Pin reads to a snapshot version so the read-dataset cache stays valid across commits that advance the table's current version (GEN-574). WARNING: INTERNAL API DETAIL

with_where_as_bool_column

with_where_as_bool_column() -> Self

Include the filter selected column in the result instead of just selected rows

to_query_object

to_query_object() -> GenevaQuery

from_query_object

from_query_object(
    table: Table, query: GenevaQuery
) -> GenevaQueryBuilder

take_rows

take_rows(rows: list[int]) -> Table

output_schema

output_schema() -> Schema

to_batches

to_batches(
    batch_size: int | None = None,
    *,
    timeout: timedelta | None = None,
) -> RecordBatchReader

to_arrow

to_arrow(*args, timeout: timedelta | None = None) -> Table

rerank

rerank(reranker: Reranker) -> Self

create_materialized_view

create_materialized_view(
    conn: Connection, view_name: str
) -> Table

Creates a materialized view of the table.

The materialized view will be a table that contains the result of the query. The view will be populated via a pipeline job.

Parameters:

• conn (Connection) –

  A connection to the database to create the view in.

• view_name (str) –

  The name of the view to create.

Raises:

• UserWarning –

  If the source table does not have stable row IDs enabled. Without stable row IDs, incremental refresh is only supported when refreshing to the same source version. Attempting to refresh to a different version will fail.