Published March 31, 2022 | Version v1
Dataset Open

Appendix of "Revisiting the Effect of Branch Handling Strategies on Change Recommendation"

  • 1. Tokyo Institute of Technology

Description

This is the dataset of the paper "Revisiting the Effect of Branch Handling Strategies on Change Recommendation", presented at ICPC 2022.

Files Overview

  • `path/to/merge_result.db`: SQLite database for the results of Imp_merge.
  • `path/to/no_merge_result.db`: SQLite database for the results of Imp_no_merge.

Database Schema

  • Table: result
    • id: ID
    • hexsha: the commit SHA-1
    • algorithm_type: the algorithm used (SINGLE or OTHERS)
    • history_type: the branch handling strategy used (FULL, FIRST_PARENTS, or BOTH). BOTH is used in case two strategies used the same commits.
    • result_type: prediction result (SUCCESS, FAILURE, NO_PREDICTION, or NO_RULE). NO_RULE is used in case of no prediction and no rule.
    • answer_file_name: the oracle filename
  • Table: prediction
    • id: ID
    • result_id: refers to result.id
    • file_name: the predicted filename
  • Table: query
    • id: ID
    • result_id: refers to result.id
    • file_name: the filename used as the query
  • Table: used_commit
    • id: ID
    • result_id: refers to result.id
    • hexsha: the commit SHA-1 to be used for recommendation
  • Table: rule
    • id: ID
    • result_id: refers to result.id
    • conclusion: the filename at the conclusion part
    • confidence: the confidence value
    • support: the support value
  • Table: rule_term (the condition part of a rule)
    • id: ID
    • rule_id: refers to rule.id
    • file_name: the filename at the condition part

Files

icpc2022-branch-handling-dataset.zip

Files (1.0 GB)

Name Size Download all
md5:089bc0af00ec4c544f947600e62aa4ee
1.0 GB Preview Download