{
  "schema_version": 2,
  "id": "integrate/redis-adbc-driver",
  "title": "Redis ADBC driver",
  "url": "https://redis.io/docs/latest/integrate/redis-adbc-driver/",
  "summary": "Query Redis with SQL and exchange data as Apache Arrow.",
  "tags": [
    "docs",
    "integrate",
    "oss",
    "rs",
    "rc"
  ],
  "last_updated": "2026-11-17T10:18:45Z",
  "children": [
    {
      "id": "integrate/redis-adbc-driver/install",
      "summary": "Install the Redis ADBC driver and check your connection to Redis.",
      "title": "Install and connect",
      "url": "https://redis.io/docs/latest/integrate/redis-adbc-driver/install/"
    },
    {
      "id": "integrate/redis-adbc-driver/quick-start",
      "summary": "Create a table, load Arrow data, and query Redis with SQL from Python.",
      "title": "Query Redis with SQL and Arrow",
      "url": "https://redis.io/docs/latest/integrate/redis-adbc-driver/quick-start/"
    }
  ],
  "page_type": "content",
  "content_hash": "a6e8da21d0ac824b48ca19c1aba89ee147a9bf30fc0aab80ec234de7ca3b85ce",
  "sections": [
    {
      "id": "overview",
      "title": "Overview",
      "role": "overview",
      "text": "The Redis Arrow Database Connectivity (ADBC) driver lets you query Redis with\nSQL and exchange data as [Apache Arrow](https://arrow.apache.org/) record\nbatches. [ADBC](https://arrow.apache.org/adbc/) is an Arrow standard for\ndatabase drivers, so any tool that can load an ADBC driver can connect to\nRedis with it. Examples include pandas, Polars, DuckDB, SQLAlchemy, Apache\nAirflow, and Dagster.\n\nThe [Redis ADBC driver project](https://github.com/redis/redis-adbc-driver)\nis open source under the Apache 2.0 license."
    },
    {
      "id": "when-to-use-the-driver",
      "title": "When to use the driver",
      "role": "content",
      "text": "To build an application, use the\n[Redis client library](https://redis.io/docs/latest/develop/clients) for your\nprogramming language. The ADBC driver is an alternative for work that fits SQL\nand Arrow, such as analytics, notebooks, dashboards, and data pipelines.\n\nThe driver supports:\n\n- SQL reads, inserts, updates, and deletes.\n- Analytical queries with filters, grouping, and aggregates.\n- Bulk loading of Arrow data.\n- Typed Arrow results, with supported filters and aggregations run by the\n  Redis [Query Engine](https://redis.io/docs/latest/develop/ai/search-and-query).\n\nThese pages show examples in Python and Go. The driver also works from other\nlanguages and from tools that load ADBC drivers. The project's\n[integration guides](https://redis.github.io/redis-adbc-driver/docs/integrations/)\ncover them.\n\nTools that can't load an ADBC driver, such as JDBC applications, can connect\nthrough the [Redis Arrow Flight server](https://redis.io/docs/latest/integrate/redis-arrow-flight)\ninstead. The server embeds its own, older copy of the driver, and the two\ndon't share a database."
    },
    {
      "id": "how-it-works",
      "title": "How it works",
      "role": "content",
      "text": "You create SQL tables through the driver. Each table has a Query Engine index,\nand the driver stores each row as a Redis hash under keys that it manages. The\nkey names are internal, so use SQL or the driver's tools to read and change\ntable data rather than reading or writing those keys directly.\n\nHashes that your applications already write don't appear as SQL tables\nautomatically. The driver's `redis-adbc` command-line tool can copy them into a\ntable or adopt them in place. See\n[Existing HASH collections](https://redis.github.io/redis-adbc-driver/docs/driver-reference/#existing-hash-collections-redis-adbc)\nin the driver reference.\n\nThe driver doesn't upgrade tables that earlier storage formats created, so a\ndriver upgrade can require you to re-create or reload tables. Check the\n[driver reference](https://redis.github.io/redis-adbc-driver/docs/driver-reference/)\nbefore you upgrade a database that holds data you need to keep."
    },
    {
      "id": "requirements",
      "title": "Requirements",
      "role": "content",
      "text": "The driver needs a Redis database with the Query Engine (Search) enabled. It\nchecks for Search when it connects and refuses to connect without it. The\nproject tests against:\n\n- Redis Open Source 8.6.2, standalone and clustered\n- Redis Software, with Redis 8.8 databases\n- Redis Cloud, with Redis 8.6.2 databases\n\nFor Redis Software and Redis Cloud databases that support it, set the\n[query performance factor](https://redis.io/docs/latest/operate/oss_and_stack/stack-with-enterprise/search/query-performance-factor)\n(QPF) to 4× and provision enough CPU for it. QPF is a performance\nrecommendation, not a connection requirement.\n\nThe driver doesn't support:\n\n- Redis Flex databases.\n- Valkey, or Redis 7.2 derivatives such as Amazon ElastiCache and Google Cloud\n  Memorystore.\n\nFor tested versions and results, see the project's\n[compatibility details](https://redis.github.io/redis-adbc-driver/docs/compatibility/)."
    },
    {
      "id": "limitations",
      "title": "Limitations",
      "role": "limits",
      "text": "- **The driver is autocommit only.** Every statement commits on its own.\n  `BEGIN` and `COMMIT` are accepted for compatibility with SQL written for\n  PostgreSQL, but they have no effect. `ROLLBACK` undoes nothing: a row that\n  you insert between `BEGIN` and `ROLLBACK` stays in the table. `SAVEPOINT` is\n  not supported.\n- **Wheels cover limited platforms.** The Python package includes the native\n  driver for Linux (x86-64 and ARM64, glibc 2.28 or later) and macOS on Apple\n  silicon (macOS 15 or later). There are no wheels for Windows, Intel Macs, or\n  musl-based Linux.\n- **Unconstrained `NUMERIC` expressions can break some readers.** A\n  calculation such as `SUM(quantity * unit_price)` returns a 256-bit Arrow\n  decimal whose precision depends on the data. Some readers, including recent\n  Polars versions, can't import it. Cast the result to a fixed precision, such\n  as `CAST(SUM(quantity * unit_price) AS NUMERIC(12,2))`.\n\nFor the complete list of supported SQL and limits, see the\n[driver reference](https://redis.github.io/redis-adbc-driver/docs/driver-reference/)."
    },
    {
      "id": "data-tools",
      "title": "Data tools",
      "role": "content",
      "text": "The project documents how to connect these tools. Each guide lists its own\nversion requirements and limitations, and some tools need custom code rather\nthan a ready-made integration.\n\n| Tool | Guide |\n| :-- | :-- |\n| pandas | [pandas guide](https://redis.github.io/redis-adbc-driver/docs/integrations/pandas/) |\n| Polars | [Polars guide](https://redis.github.io/redis-adbc-driver/docs/integrations/polars/) |\n| SQLAlchemy | [SQLAlchemy guide](https://redis.github.io/redis-adbc-driver/docs/integrations/sqlalchemy/) |\n| Apache Superset | [Superset guide](https://redis.github.io/redis-adbc-driver/docs/integrations/superset/) |\n| DuckDB | [DuckDB guide](https://redis.github.io/redis-adbc-driver/docs/duckdb/) |\n| Apache Airflow | [Airflow guide](https://redis.github.io/redis-adbc-driver/docs/integrations/airflow/) |\n| Dagster | [Dagster guide](https://redis.github.io/redis-adbc-driver/docs/integrations/dagster/) |\n| Prefect | [Prefect guide](https://redis.github.io/redis-adbc-driver/docs/integrations/prefect/) |\n| dlt | [dlt guide](https://redis.github.io/redis-adbc-driver/docs/integrations/dlt/) |\n\nThe project's\n[ecosystem comparison](https://redis.github.io/redis-adbc-driver/docs/data-engineering-adbc-support/)\ncovers about 50 tools, including which connection paths are tested and which\nneed custom work."
    },
    {
      "id": "next-steps",
      "title": "Next steps",
      "role": "content",
      "text": "- [Install and connect](https://redis.io/docs/latest/integrate/redis-adbc-driver/install)\n- [Query Redis with SQL and Arrow](https://redis.io/docs/latest/integrate/redis-adbc-driver/quick-start)"
    }
  ],
  "examples": []
}
