The author explores the use of Apache DataFusion for writing graph algorithms, specifically focusing on the weakly connected components (WCC) problem. They highlight their background as an inexperienced Rust developer who is new to DataFusion yet keen on graph algorithms. The discussion centers on how WCC relates to identity resolution in modern data warehouses (DWHs), where disparate data sources often lack a unified ID system. The author aims to tackle this challenge by treating IDs and their attributes as vertices in a graph, with edges representing matching attributes. This approach allows for the identification of connected components, ultimately leading to the creation of a “golden row” or “super-ID” that consolidates multiple IDs into a single entity. The author provides a brief overview of graph theory concepts necessary for understanding the weakly connected components problem. They explain the structure of a graph, the definition of connected components, and the distinction between weakly and strongly connected components. The article emphasizes that while many graph problems are niche, the connected components problem frequently intersects with the work of data engineers, especially in identity resolution scenarios. The author describes how traditional SQL methods for resolving identities can be inefficient, particularly when dealing with large datasets. The classical algorithm for finding connected components involves using Breadth-First Search (BFS) to explore the graph. The algorithm iterates over vertices, identifying and grouping connected vertices into components. The author shares a code snippet from the NetworkX library, illustrating how the algorithm operates. This practical approach to implementing the WCC algorithm with DataFusion provides a foundation for further exploration and optimization in graph processing, especially within the context of data warehousing challenges.