GDED (GitDelver Enterprise Dataset)

This dataset was produced by analyzing the Git commits from 101 closed-source software repositories with the GitDelver mining tool (Zenodo, GitHub).

It is made of three CSV files.

commits_history.csv shows data about 106,216 Git commits performed by 164 developers. It has the following columns:

Repository: the name of the repository.

Branches: the list of branches in which this modification has been integrated (works best if you target a bare repository).

NbBranches: the number of branches in which this modification has been integrated (works best if you target a bare repository).

CommitId: the identifier of the commit.

Author: the author of the modification.

DateTime: the date and time of the modification.

Date: the date of the modification.

HourOfDay: the hour of the day at which the modification took place.

Merge: flag telling if the commit is a merge commit.

BugFix: flag telling if the modification is a bugfix.

SATD: flag telling if the modification contains Self-Admitted Technical Debt.

NbModifiedFiles: the total number of files (supported and unsupported) modified by this commit.

ModifiedFiles: the list of files modified by this commit.

NbModifiedProdSourceFiles: the number of production source files modified by this commit.

NbModifiedTestSourceFiles: the number of test source files modified by this commit.

NbModifications: the total number of modifications done by this commit.

NbInsertions: the number of insertions done by the commit.

NbDeletions: the number of deletions done by the commit.

files_history.csv shows data about 470,940 file modifications performed by 153 developers. It has the following columns:

Repository: the name of the repository.

Branches: the list of branches in which this modification has been integrated (works best if you target a bare repository).

NbBranches: the number of branches in which this modification has been integrated (works best if you target a bare repository).

OldFilePath: the old relative path to the file.

FilePath: the relative path to the file.

FileName: the name of the file.

FileExtension: the file extension.

FileType: the type of the file ("Production" or "Test").

ChangeType: the type of the change ("ADD", "COPY", "RENAME", "DELETE", "MODIFY" or "UNKNOWN").

NbMethods: the number of methods in the file.

NbMethodsChanged: the number of methods that have been modified in this file for this commit.

NLOC: the number of lines of code of the file.

Complexity: the Weighted Methods per Class complexity, i.e., the sum of the cyclomatic complexity numbers of all the methods of the file.

NlocDivByNbMethods: the number of lines of code of the file divided by the number of methods of the file.

ComplexDivByNbMethods: the complexity of the file divided by the number of methods of the file.

SATD: flag telling if the modification contains Self-Admitted Technical Debt.

NbLinesAdded: the number of lines added.

NbLinesDeleted: the number of lines deleted.

CommitId: the identifier of the commit.

Author: the author of the modification.

DateTime: the date and time of the modification.

Date: the date of the modification.

HourOfDay: the hour of the day at which the modification took place.

methods_history.csv shows data about 3,471,556 method modifications performed by 153 developers. It has the following columns:

Repository: the name of the repository.

Branches: the list of branches in which this modification has been integrated (works best if you target a bare repository).

NbBranches: the number of branches in which this modification has been integrated (works best if you target a bare repository).

OldFilePath: the old relative path to the file.

FilePath: the relative path to the file.

FileName: the name of the file.

FileType: the type of the file ("Production" or "Test").

MethodName: the name of the method.

NbParams: the number of parameters in the method signature.

NLOC: the number of lines of code of the method.

Complexity: the cyclomatic complexity number of the method.

CommitId: the identifier of the commit.

Author: the author of the modification.

DateTime: the date and time of the modification.

Date: the date of the modification.

HourOfDay: the hour of the day at which the modification took place.

Remarks:

All of the data come from a single organization in the industry. It wishes to remain anonymous.

files_history.csv references commits_history.csv and methods_history.csv references files_history.csv and commits_history.csv.

For privacy (GDPR) and security reasons, several columns have been anonymized or have had their contents scrambled.