MagicStack/asyncpg
Asyncpg
asyncpg is an asynchronous database driver and binary protocol client for PostgreSQL. It provides a non-blocking interface for executing SQL statements, streaming result sets, and managing data transfer between an application and a PostgreSQL database.
The driver implements the PostgreSQL binary protocol directly to facilitate efficient data transfer and type conversion. It includes a connection pool to maintain and reuse open database connections, reducing the latency associated with repeated handshakes.
The project covers a broad range of database integration capabilities, including atomic transaction management and the mapping of standard, composite, and custom data types to native objects. It further optimizes data retrieval through the use of server-side prepared statements and cursor-based result streaming to manage memory usage.
Features
- Asynchronous Database Drivers - Serves as a high-performance, non-blocking PostgreSQL client designed for asynchronous data operations.
- Binary Protocol Implementations - Directly implements the PostgreSQL binary wire protocol for high-performance asynchronous data transfer.
- Asynchronous Database Command Execution - Executes database commands through a non-blocking interface to maintain application responsiveness.
- Binary Type Codecs - Uses binary-format codecs to convert database types to Python objects, reducing parsing overhead.
- Composite Type Encoders - Encodes and decodes composite types, arrays, and custom formats between the database and application.
- PostgreSQL Pool Managers - Provides a connection pool specifically optimized for PostgreSQL wire protocol and server behavior.
- PostgreSQL Integrations - Provides a high-performance asynchronous integration layer for PostgreSQL relational database connectivity.
- Data Type Mappings - Translates complex database structures like arrays and records into native Python collections.
- Atomic Transactions - Ensures data integrity by grouping multiple database operations into single atomic units of work.
- Database Transaction Management - Groups SQL operations into atomic units of work to ensure data integrity and consistency.
- Connection Pooling - Maintains a pool of reusable active connections to eliminate the latency of repeated handshakes.
- Query Result Fetching - Retrieves resulting data sets from SQL statements using asynchronous communication.
- SQL Query Interfaces - Provides a non-blocking programming interface for executing SQL statements and streaming result sets.
- Coroutine-Based Asynchronous I/O - Implements a non-blocking interface using coroutines and an event loop for high-concurrency database communication.
- Database Connectivity Drivers - Establishes secure, asynchronous links to PostgreSQL servers using connection strings.
- Connection Pools - Uses connection pooling to maintain a cache of reusable database links for high-traffic applications.
- Nested Transactions - Supports nested work units through transaction blocks and savepoints.
- Memory-Optimized Data Retrieval - Implements cursor-based result streaming to retrieve large datasets while minimizing application memory overhead.
- Prepared Statement Interfaces - Supports server-side prepared statements to pre-compile queries and improve execution speed.
- Prepared Statements - Utilizes server-side prepared statements to pre-compile queries for increased execution speed and security.
- Result Streaming - Retrieves large datasets incrementally using cursors to prevent application memory exhaustion.
- Database Drivers - High-performance PostgreSQL database client library.
