Distributed query engine providing simple and reliable data processing for any modality and scale

Website • Docs • Installation • Daft Quickstart • Community and Support

Daft is a distributed query engine for large-scale data processing using Python or SQL, implemented in Rust.

• Familiar interactive API: Lazy Python Dataframe for rapid and interactive iteration, or SQL for analytical queries
• Focus on the what: Powerful Query Optimizer that rewrites queries to be as efficient as possible
• Data Catalog integrations: Full integration with data catalogs such as Apache Iceberg
• Rich multimodal type-system: Supports multimodal types such as Images, URLs, Tensors and more
• Seamless Interchange: Built on the Apache Arrow In-Memory Format
• Built for the cloud: Record-setting I/O performance for integrations with S3 cloud storage

Table of Contents

• About Daft
• Getting Started
• Benchmarks
• Contributing
• Telemetry
• Related Projects
• License

Daft was designed with the following principles in mind:

1. Any Data: Beyond the usual strings/numbers/dates, Daft columns can also hold complex or nested multimodal data such as Images, Embeddings and Python objects efficiently with its Arrow based memory representation. Ingestion and basic transformations of multimodal data is extremely easy and performant in Daft.
2. Interactive Computing: Daft is built for the interactive developer experience through notebooks or REPLs - intelligent caching/query optimizations accelerates your experimentation and data exploration.
3. Distributed Computing: Some workloads can quickly outgrow your local laptop's computational resources - Daft integrates natively with Ray for running dataframes on large clusters of machines with thousands of CPUs/GPUs.

Install Daft with pip install daft.

For more advanced installations (e.g. installing from source or with extra dependencies such as Ray and AWS utilities), please see our Installation Guide

Check out our quickstart!

In this example, we load images from an AWS S3 bucket's URLs and resize each image in the dataframe:

import daft

# Load a dataframe from filepaths in an S3 bucket
df = daft.from_glob_path("s3://daft-public-data/laion-sample-images/*")

# 1. Download column of image URLs as a column of bytes
# 2. Decode the column of bytes into a column of images
df = df.with_column("image", df["path"].url.download().image.decode())

# Resize each image into 32x32
df = df.with_column("resized", df["image"].image.resize(32, 32))

df.show(3)

To see the full benchmarks, detailed setup, and logs, check out our benchmarking page.

• Daft Quickstart - learn more about Daft's full range of capabilities including dataloading from URLs, joins, user-defined functions (UDF), groupby, aggregations and more.
• User Guide - take a deep-dive into each topic within Daft
• API Reference - API reference for public classes/functions of Daft
• SQL Reference - Daft SQL reference

We <3 developers! To start contributing to Daft, please read CONTRIBUTING.md. This document describes the development lifecycle and toolchain for working on Daft. It also details how to add new functionality to the core engine and expose it through a Python API.

Here's a list of good first issues to get yourself warmed up with Daft. Comment in the issue to pick it up, and feel free to ask any questions!

To help improve Daft, we collect non-identifiable data via Scarf (https://scarf.sh).

To disable this behavior, set the environment variable DO_NOT_TRACK=true.

The data that we collect is:

1. Non-identifiable: Events are keyed by a session ID which is generated on import of Daft
2. Metadata-only: We do not collect any of our users’ proprietary code or data
3. For development only: We do not buy or sell any user data

Please see our documentation for more details.

Check out our engine comparison page for more details!

Daft has an Apache 2.0 license - please see the LICENSE file.