Sign in to view source links and access this dataset
Description
Victor Cacciari Miraldo from Utrecht University collected a dataset of merge conflicts from GitHub repositories. The data includes files representing the original code, two conflicting versions, and the final merged solution committed by a developer. The repositories were selected by manually searching for a programming language and choosing those with many forks, commits, and contributors.
Use Cases
Train machine learning models for automatic merge conflict resolution based on the provided conflict and solution pairs.
Analyze patterns in how developers resolve merge conflicts based on the committed solutions.
Benchmark the performance of automated software merging tools against real-world developer resolutions.
Strengths
Includes the developer's final merged solution for each conflict, providing ground truth.
Repositories were selected based on activity metrics like number of forks, commits, and contributors.
Limitations
Row count and dataset size are unknown, which may limit suitability assessment.
Column-level documentation is absent; field semantics must be inferred after download.
Last update date is unknown; freshness unverified.
Provenance
Source
Utrecht University
Collection Method
Manually collected from GitHub repositories based on programming language and repository activity.
License is listed as Open Access (green); specific terms should be verified from the source.