Download
SQLFlow is an open source project designed to bridge the gap between traditional SQL-based data processing and modern machine learning workflows by extending SQL syntax with AI capabilities. It acts as a compiler that translates SQL programs into executable workflows, enabling users to train, evaluate, and deploy machine learning models directly from SQL statements. It integrates with multiple database engines such as MySQL, Hive, and MaxCompute, while also supporting machine learning frameworks like TensorFlow and XGBoost. By embedding machine learning operations into SQL, it removes the need for users to switch between programming languages such as Python or R, simplifying the overall workflow. SQLFlow also supports model training, prediction, and explanation tasks, allowing data practitioners to work entirely within a familiar query interface.
Features
- Extends SQL syntax to support machine learning tasks like training and inference
- Compiles SQL programs into Kubernetes-native workflows for execution
- Integrates with databases such as MySQL, Hive, and MaxCompute
- Supports ML frameworks including TensorFlow and XGBoost
- Enables model training, prediction, and explanation directly in SQL
- Reduces need for separate programming languages in ML pipelines
Project Samples
