Trino

Trino federation connector.

Trino-dialect SQL helpers.

Trino is the OSS query engine commonly paired with the open-source unitycatalog.io implementation when callers need live federation to a foreign RDBMS — OSS UC itself has no native federation primitive (see SU#519 / SU#521).

The first surface here is siege_utilities.trino.federation, which generates CREATE VIEW statements that select through Trino’s postgresql connector. This is the OSS analog of siege_utilities.databricks.lakehouse_federation for the Databricks Lakehouse Federation use case.

siege_utilities.trino.build_trino_federation_view_sql(target_catalog, target_schema, target_view, postgres_catalog, source_schema, source_table=None)[source]

Build SQL for a Trino federation view over a PostgreSQL source.

Generates a CREATE OR REPLACE VIEW statement in target_catalog.target_schema.target_view that selects every column from the PostgreSQL table postgres_catalog.source_schema.source_table via Trino’s postgresql connector.

All identifier fields are validated against the SU identifier allow-list before interpolation. The output is double-quoted in Trino’s ANSI dialect (not backticks).

Parameters:

  • target_catalog (str) – Trino catalog the view lives in (e.g. "iceberg" for an OSS UC-backed Iceberg catalog).

  • target_schema (str) – Schema within the target catalog (e.g. "analytics").

  • target_view (str) – View name (e.g. "persons").

  • postgres_catalog (str) – Trino catalog alias for the PostgreSQL connector (whatever the operator named the etc/catalog/postgresql.properties file, sans extension).

  • source_schema (str) – PostgreSQL schema (e.g. "public").

  • source_table (str | None) – PostgreSQL table name. Defaults to target_view when omitted, which is the common case (same name on both sides).

Returns:

A single SQL statement ending in ;.

Return type:

str

Example

>>> build_trino_federation_view_sql(
...     target_catalog="iceberg",
...     target_schema="analytics",
...     target_view="persons",
...     postgres_catalog="postgresql",
...     source_schema="public",
... )
'CREATE OR REPLACE VIEW "iceberg"."analytics"."persons" AS\nSELECT * FROM "postgresql"."public"."persons";'
siege_utilities.trino.build_trino_schema_and_view_sync_sql(target_catalog, target_schema, postgres_catalog, source_schema, tables)[source]

Build CREATE SCHEMA + per-table CREATE VIEW statements.

Convenience for syncing many PostgreSQL tables into a Trino catalog in one pass. The schema statement uses IF NOT EXISTS so the list is safe to re-run. Each view statement is CREATE OR REPLACE so re-runs after upstream column additions stay correct.

Parameters:

  • target_catalog (str) – Trino catalog the schema + views live in.

  • target_schema (str) – Schema within target_catalog.

  • postgres_catalog (str) – Trino PostgreSQL connector alias.

  • source_schema (str) – PostgreSQL schema containing tables.

  • tables (Iterable[str]) – Iterable of PostgreSQL table names. Same names are used for the Trino views.

Returns:

one CREATE SCHEMA followed by one CREATE OR REPLACE VIEW per table, in the order given.

Return type:

List of SQL statements

siege_utilities.trino.quote_ident(value)[source]

Quote an identifier in Trino dialect.

Trino uses ANSI double quotes for identifiers (not backticks like Databricks). Internal double-quote characters are escaped by doubling. Callers must validate identifiers before passing them here (the higher-level federation builders call validate_sql_identifier before quote_ident).

Parameters:

value (str)

Return type:

str

Submodules

Trino SQL helpers for federating to a foreign RDBMS via Trino connectors.

This is the OSS analog of siege_utilities.databricks.lakehouse_federation for the open-source unitycatalog.io (and any other catalog Trino can write a VIEW into).

OSS Unity Catalog does NOT support live query federation natively (no CREATE FOREIGN TABLE/USING CONNECTION syntax — that is Databricks-specific). The OSS path is to configure Trino with the postgresql connector (catalog config file), then expose the foreign table under a stable name in a different Trino catalog via CREATE OR REPLACE VIEW. Readers query the view; Trino does the federation transparently.

This module generates the view-creation SQL. Catalog configuration (the etc/catalog/*.properties files Trino reads at startup) is out of scope — that is operator territory.

See SU#521 (umbrella) and SU#519 (root context).

siege_utilities.trino.federation.build_trino_federation_view_sql(target_catalog, target_schema, target_view, postgres_catalog, source_schema, source_table=None)[source]

Build SQL for a Trino federation view over a PostgreSQL source.

Generates a CREATE OR REPLACE VIEW statement in target_catalog.target_schema.target_view that selects every column from the PostgreSQL table postgres_catalog.source_schema.source_table via Trino’s postgresql connector.

All identifier fields are validated against the SU identifier allow-list before interpolation. The output is double-quoted in Trino’s ANSI dialect (not backticks).

Parameters:

  • target_catalog (str) – Trino catalog the view lives in (e.g. "iceberg" for an OSS UC-backed Iceberg catalog).

  • target_schema (str) – Schema within the target catalog (e.g. "analytics").

  • target_view (str) – View name (e.g. "persons").

  • postgres_catalog (str) – Trino catalog alias for the PostgreSQL connector (whatever the operator named the etc/catalog/postgresql.properties file, sans extension).

  • source_schema (str) – PostgreSQL schema (e.g. "public").

  • source_table (str | None) – PostgreSQL table name. Defaults to target_view when omitted, which is the common case (same name on both sides).

Returns:

A single SQL statement ending in ;.

Return type:

str

Example

>>> build_trino_federation_view_sql(
...     target_catalog="iceberg",
...     target_schema="analytics",
...     target_view="persons",
...     postgres_catalog="postgresql",
...     source_schema="public",
... )
'CREATE OR REPLACE VIEW "iceberg"."analytics"."persons" AS\nSELECT * FROM "postgresql"."public"."persons";'
siege_utilities.trino.federation.build_trino_schema_and_view_sync_sql(target_catalog, target_schema, postgres_catalog, source_schema, tables)[source]

Build CREATE SCHEMA + per-table CREATE VIEW statements.

Convenience for syncing many PostgreSQL tables into a Trino catalog in one pass. The schema statement uses IF NOT EXISTS so the list is safe to re-run. Each view statement is CREATE OR REPLACE so re-runs after upstream column additions stay correct.

Parameters:

  • target_catalog (str) – Trino catalog the schema + views live in.

  • target_schema (str) – Schema within target_catalog.

  • postgres_catalog (str) – Trino PostgreSQL connector alias.

  • source_schema (str) – PostgreSQL schema containing tables.

  • tables (Iterable[str]) – Iterable of PostgreSQL table names. Same names are used for the Trino views.

Returns:

one CREATE SCHEMA followed by one CREATE OR REPLACE VIEW per table, in the order given.

Return type:

List of SQL statements

siege_utilities.trino.federation.quote_ident(value)[source]

Quote an identifier in Trino dialect.

Trino uses ANSI double quotes for identifiers (not backticks like Databricks). Internal double-quote characters are escaped by doubling. Callers must validate identifiers before passing them here (the higher-level federation builders call validate_sql_identifier before quote_ident).

Parameters:

value (str)

Return type:

str