sqlspec

Type-safe SQL queries across multiple databases without ORM overhead are executed using dialect-aware adapters that handle parameter style conversion and connection lifecycle automatically, supporting 18+ backends including PostgreSQL, MySQL, Oracle, BigQuery, and DuckDB.

Yes, you can integrate DuckDB analytics with a PostgreSQL primary store by utilizing multi-adapter support and native Arrow transfer to build cross-database data pipelines for analytics workloads.

Parameterized paginated queries with ordering and search filters in a Litestar project are executed by configuring an asyncpg adapter, defining a typed dataclass model, and leveraging first-party framework plugins with built-in dependency injection.

Yes, bulk data ingest and Apache Arrow integration are supported for analytics pipelines through a storage bridge and native ADBC integration, enabling efficient cross-database data transfer without ORM constraints.

First-party plugin integrations for database connection lifecycle management are available for Litestar, FastAPI, Starlette, and Flask, providing built-in dependency injection, commit modes, session stores, and filter dependencies.

You should avoid a traditional ORM for multi-database Python applications when you require native database feature access like vector search and event channels, dialect-aware transpilation across 18+ adapters, or zero lock-in for analytics pipelines.