Getting Started
Install
Install the sqlload command from PyPI, as a Python package or a standalone executable.
SQLLoad installs from PyPI. Install it as a tool with uv:
uv tool install sqlloadsqlload --version$ sqlload --versionsqlload 0.32.0Choose a package
Section titled “Choose a package”Two packages provide the same sqlload command. Install only one of them, because both provide a
sqlload command.
| Package | What it is | Python sources |
|---|---|---|
sqlload |
A native Python extension. Each platform has one wheel, built for Python’s stable ABI (abi3), which runs on every CPython from 3.10. Its sqlload command runs the loader in the same process as your Python sources. |
Yes |
sqlload-cli |
The same command as a standalone executable, without Python. | No |
Use sqlload if any table is defined in Python (Python sources). Use
sqlload-cli where Python isn’t available and every source is declared in sqlload.toml.
uv tool install sqlload-cliWith pip, in a virtual environment:
pip install sqlloadBoth packages are released for:
| Platform | Architectures |
|---|---|
| Linux (glibc 2.28 or newer) | x86_64, aarch64 |
| macOS 11 or newer | arm64 |
Use it from Python
Section titled “Use it from Python”The sqlload package also imports as a library, for declaring Python sources:
import sqlload as sl
print(sl.__version__)print(sl.executable()) # the sqlload command installed alongside the packageSee the Python API.
Follow the Quickstart to copy a folder into DuckDB and load a Postgres database with a project.