Delta Arrow Reader

Delta Arrow Reader lets you read Delta Lake tables from Python, Rust, or the command line. Select the rows and columns you need, then process them in batches instead of loading the whole result into memory. Each batch is a group of rows in Apache Arrow format, ready to use in your application.

The reader works with existing Delta tables. It does not write tables or manage transactions. Queries run on your machine or in your application's environment, without a Spark or Databricks cluster. Cloud tables require direct storage access.

Start here

Choose how you want to read your table. Each quickstart links to the installation steps for that interface:

• Python: read rows with PyArrow or query them with SQL.
• Command line: inspect a table or save rows to a file.
• Rust: process Arrow batches in your application.
• DataFusion in Rust: register a table and query it with SQL.

After your first read

• Troubleshoot a failed read in Python, the command line, or Rust.
• Connect to S3 or Azure, or adjust opening time and resource use.
• Learn how the reader works.
• Review the performance reports and their test conditions.
• Look up a type or method in the Python API reference or Rust API reference.