Dune Client

A python framework for interacting with Dune Analytics' officially supported API service.

Installation

Using uv (recommended)

uv is a fast Python package and project manager written in Rust. It provides:

• ⚡ 10-100x faster dependency resolution
• 🔒 Reproducible builds with lockfiles
• 🐍 Automatic Python version management
• 📦 Zero-configuration virtual environments

Install uv first, then:

# Add to an existing project
uv add dune-client

# Or create a new project
uv init my-dune-project
cd my-dune-project
uv add dune-client

Using pip

pip install dune-client

Example Usage

Quickstart: run_query

Export your DUNE_API_KEY (or place it in a .env file - as in here .env.sample and source .env).

💡 Tip: If using uv, you can run examples directly without activating a virtual environment:

uv run python your_script.py

from dune_client.types import QueryParameter
from dune_client.client import DuneClient
from dune_client.query import QueryBase

query = QueryBase(
    name="Sample Query",
    query_id=1215383,
    params=[
        QueryParameter.text_type(name="TextField", value="Word"),
        QueryParameter.number_type(name="NumberField", value=3.1415926535),
        QueryParameter.date_type(name="DateField", value="2022-05-04 00:00:00"),
        QueryParameter.enum_type(name="ListField", value="Option 1"),
    ],
)
print("Results available at", query.url())

dune = DuneClient()
results = dune.run_query(query)

# or as CSV
# results_csv = dune.run_query_csv(query)

# or as Pandas Dataframe
# results_df = dune.run_query_dataframe(query)

Further Examples

Get Latest Results

Use get_latest_results to get the most recent query results without using execution credits. You can specify a max_age_hours to re-run the query if the data is too outdated.

from dune_client.client import DuneClient

dune = DuneClient()
results = dune.get_latest_result(1215383, max_age_hours=8)

Paid Subscription Features

CRUD Operations

If you're writing scripts that rely on Dune query results and want to ensure that your local, peer-reviewed, queries are being used at runtime, you can call update_query before run_query!

Here is a fictitious example making use of this functionality;

from dune_client.types import QueryParameter
from dune_client.client import DuneClient

sql = """
    SELECT block_time, hash,
    FROM ethereum.transactions
    ORDER BY CAST(gas_used as uint256) * CAST(gas_price AS uint256) DESC
    LIMIT {{N}}
    """

dune = DuneClient()
query = dune.create_query(
    name="Top {N} Most Expensive Transactions on Ethereum",
    query_sql=sql,
    # Optional fields
    params=[QueryParameter.number_type(name="N", value=10)],
    is_private=False,  # default
)
query_id = query.base.query_id
print(f"Created query with id {query.base.query_id}")
# Could retrieve using
# dune.get_query(query_id)

dune.update_query(
    query_id,
    # All parameters below are optional
    name="Top {N} Most Expensive Transactions on {Blockchain}",
    query_sql=sql.replace("ethereum", "{{Blockchain}}"),
    params=query.base.parameters() + [QueryParameter.text_type("Blockchain", "ethereum")],
    description="Shows time and hash of the most expensive transactions",
    tags=["XP€N$IV $H1T"],
)

dune.archive_query(query_id)
dune.unarchive_query(query_id)

dune.make_private(query_id)
dune.make_public(query_id)

Developer Usage & Deployment

This project uses uv for dependency management and development workflows.

Setup

# Clone the repository
git clone https://github.com/duneanalytics/dune-client.git
cd dune-client

# Install dependencies
uv sync --dev

Development Commands

# Format code
uv run ruff format

# Lint code
uv run ruff check

# Type checking
uv run mypy dune_client/ --strict

# Run tests
uv run python -m pytest tests/unit      # Unit tests
uv run python -m pytest tests/e2e       # E2E tests (requires DUNE_API_KEY)

Multi-Python Testing

This project supports Python 3.11, 3.12, and 3.13. You can test across all versions using tox with uv:

# Test all Python versions with tox
uv run tox
# or
make test-tox

# Test specific Python versions
uv run tox -e py311
uv run tox -e py312
uv run tox -e py313

Makefile Shortcuts

Installation

make install        # Uses uv sync
# or
make install-dev    # Uses uv sync --dev

Format, Lint & Types

make check

can also be run individually with fmt, lint and types respectively.

Testing

make test-unit  # Unit tests
make test-e2e   # Requires valid `DUNE_API_KEY`
make test-all   # Both unit and e2e tests
make test-tox   # Multi-Python testing (py311, py312, py313)

Deployment

Publishing releases to PyPi is configured automatically via github actions (cf. ./.github/workflows/py-publish.yaml). Any time a branch is tagged for release this workflow is triggered and published with the same version name.