What problem does it solve?
Setting up MATLAB to work with Databricks involves choosing between Spark and JDBC paths, configuring authentication, matching Python and Java versions, and avoiding common pitfalls like pulling unfiltered large tables or building fragile JDBC URLs manually. This Skill guides the entire connection workflow so data moves correctly between MATLAB and Databricks.
Core Features & Use Cases
- Dual Connection Paths: Create Spark sessions with getDatabricksSession() for server-side filtering of large tables, or JDBC connections with databricks.JDBCConnection and StandaloneJDBCConnection for SQL-based workflows via Database Toolbox.
- Authentication Setup: Configure OauthU2M, OauthM2M, or PAT authentication through .databrickscfg profiles, environment variables, or token passthrough, including serverless compute and on-Databricks browser scenarios.
- Data Operations: Read Unity Catalog tables and Volume files, filter and transform DataFrames server-side, and write results back using Spark DataFrames or JDBC sqlwrite with Simba driver optimization.
- Use Case: An engineer needs to analyze millions of sensor readings stored in Unity Catalog. The Skill guides them to create a Spark session, filter rows server-side by date and temperature thresholds, and pull only the filtered results into a MATLAB table for analysis.
Quick Start
Ask your AI agent to connect MATLAB to your Databricks workspace and read a filtered subset of a Unity Catalog table into a MATLAB table.