dbc

Overview

dbc is a command-line tool for installing and managing ADBC drivers.

dbc can:

• Install pre-built ADBC drivers with a single command
• Install drivers in your user account, on the system, or in virtual environments
• Manage isolated, reproducible project environments with driver lists and lockfiles
• Run on macOS, Linux, and Windows
• Be installed in many ways (with pip, standalone installers, Docker images, and more)
• Work in CI/CD environments

Installation

There are multiple ways to install dbc:

From PyPI

For simple installation, we recommend the popular pipx tool which will automatically put it on your PATH:

pipx install dbc

You can also just test it out instead of installing it:

pipx run dbc

You can also use a virtual environment:

python -m venv .venv
source .venv/bin/activate
pip install dbc

Standalone Installer

macOS or Linux

You can download the install script and execute it:

curl -LsSf https://dbc.columnar.tech/install.sh | sh

If your system doesn't have curl you can also use wget:

wget -q0- https://dbc.columnar.tech/install.sh | sh

If you want to inspect the script before use, you can simply run:

curl -LsSf https://dbc.columnar.tech/install.sh | less

Windows

Download the Windows graphical installer for your architecture:

Architecture	Installer
x64 (64-bit)	https://dbc.columnar.tech/latest/dbc-latest-x64.msi
x86 (32-bit)	https://dbc.columnar.tech/latest/dbc-latest-x86.msi
ARM64	https://dbc.columnar.tech/latest/dbc-latest-arm64.msi

Or use irm to download the install script and execute it with iex:

powershell -ExecutionPolicy ByPass -c "irm https://dbc.columnar.tech/install.ps1 | iex

Changing the execution policy allows running a script from the internet.

Of course, you can also inspect the script before use:

powershell -c "irm https://dbc.columnar.tech/install.ps1 | more"

GitHub Releases

Release artifacts for dbc can also be downloaded directly from GitHub Releases. Included in the artifacts are also cryptographic signatures and a checksum file to ensure nothing has been tampered with.

Each release includes the following assets allowing you to install using your preferred method:

• .tar.gz or .zip archives containing the appropriate binary for all supported platforms and architectures
• .deb and .rpm installation packages
• An msi installer package for Windows
• .snap packages
• Python wheel packages that bundle the dbc executable binary

Docker

Docker images are also provided with standalone binaries that can be easily run using:

docker run --rm -it columnar/dbc:latest --help

Available Images

The following distroless images are available for linux-based amd64 and arm64 architectures:

• columnar/dbc:latest
• columnar/dbc:{major}.{minor}.{patch}, e.g. columnar/dbc:0.0.1

Getting Started

Once you have dbc available to you on the command line, you can install an ADBC driver and make it available to your user. For example, to install the bigquery driver:

dbc install bigquery

Alternately, when working on a project you can create a dbc.toml file to create a list of drivers to install to create a reproducible environment:

cd <path/to/project>
dbc init # creates dbc.toml
dbc add bigquery # adds this to the driver list
dbc sync # install drivers and create dbc.lock

Using dbc add also allows version constraints:

dbc add "bigquery>=1.0.0"
dbc sync # looks for and installs a version >=1.0.0

Using the Driver

The simplest way to use the driver is via Python with adbc-driver-manager. Note: version 1.8.0 added support for driver manifests, so you'll need that version of the driver manager or higher.

dbc install snowflake
pip install "adbc-driver-manager>=1.8.0"

Using the driver is easy:

import adbc_driver_manager.dbapi as adbc

snowflake_connect_args = {
    "username": "USER",
    "password": "PASS",
    "adbc.snowflake.sql.account": "ACCOUNT-IDENT",
    "adbc.snowflake.sql.db": "SNOWFLAKE_SAMPLE_DATA",
    # other connect options
}

with adbc.connect(
    driver="snowflake",
    db_kwargs=snowflake_connect_args,
) as con, con.cursor() as cursor:
    cursor.execute("SELECT * FROM CUSTOMER LIMIT 5")
    table = cursor.fetch_arrow_table()

print(table)

For more examples, see the documentation!

Communications

For general questions and discussion, use the GitHub discussions.

To report an issue, request a feature, or contribute an improvement, use the GitHub issues and PRs.