arrow-odbc

What it is and what it does

arrow-odbc bridges ODBC data sources and Apache Arrow by reading SQL query results directly into Arrow record batches. It wraps a Rust crate and uses cffi to bind Python with efficient bulk-read operations, avoiding serialization overhead. The package handles type mapping between ODBC and Arrow schemas (e.g., SQL Numeric to Decimal128, Timestamp to TimestampNanosecond) and supports parameterized queries.

Typical workflows involve querying any ODBC-accessible database—MySQL, MS SQL, Excel, etc.—and receiving data as Arrow batches that can be converted to pandas DataFrames or processed in columnar form. It also supports inserting Arrow data back into tables. The package requires a system-level ODBC driver manager but provides prebuilt wheels for macOS (Intel and ARM), Linux, and Windows, making deployment straightforward once the driver manager is present.

Use it for:

• Extract data from SQL Server, MySQL, or other ODBC sources into Arrow for zero-copy analytics pipelines.
• Bulk-insert pandas DataFrames or Arrow tables into databases via ODBC without intermediate serialization.
• Read Excel spreadsheets through ODBC drivers and convert to Arrow for data processing.
• Build ETL workflows that move data between legacy ODBC sources and modern columnar storage.
• Integrate ODBC data access into data science notebooks with efficient memory usage via Arrow batches.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Reads data from ODBC data sources directly into Apache Arrow record batches, enabling efficient columnar data access from databases, spreadsheets, and other ODBC-compatible sources.

Yes, if you need to read from ODBC data sources and want efficient columnar access. The MIT license, active maintenance, and prebuilt wheels make it low-friction for most platforms. Install friction is moderate due to the system ODBC driver manager requirement, but this is a one-time setup. No known vulnerabilities. Suitable for production use in data pipelines and analytics workflows.

Install

arrow-odbc on PyPI

pip install arrow-odbc

uv add arrow-odbc

poetry add arrow-odbc

Installing arrow-odbc

Before you install

Medium install friction due to system-level ODBC driver manager dependency (unixodbc-dev on Linux, UnixODBC via homebrew on macOS, preinstalled on Windows). Package provides prebuilt wheels for common platforms but requires external system library. Actively maintained with recent releases.

License in practice

MIT license permits commercial and private use with minimal restrictions, making it suitable for most projects without licensing concerns.

Quickstart

pip install arrow-odbc

from arrow_odbc import connect

connection = connect(
    connection_string="Driver={ODBC Driver 18 for SQL Server};Server=localhost;",
    user="SA",
    password="password",
)
reader = connection.read_arrow_batches(query="SELECT * FROM MyTable")
for batch in reader:
    df = batch.to_pandas()

System ODBC driver manager must be installed (unixodbc-dev on Ubuntu, UnixODBC on macOS via homebrew, preinstalled on Windows). Requires Python >= 3.10.

Verify before relying

• Performance characteristics compared to direct database drivers or other Arrow connectors in typical workloads.
• Completeness of ODBC driver support across different database systems and edge cases.
• Stability and maturity level for production use in high-throughput scenarios.

Package facts

Evidence: arrow_odbc-10.4.2-py3-none-macosx_10_12_x86_64.whl; arrow_odbc-10.4.2-py3-none-macosx_11_0_arm64.whl; arrow_odbc-10.4.2-py3-none-manylinux_2_28_aarch64.whl; arrow_odbc-10.4.2-py3-none-manylinux_2_28_x86_64.whl; arrow_odbc-10.4.2-py3-none-win_amd64.whl

Tags

More Database packages