Data Schema

Data Schema

This page describes the data schema of our merge conflict dataset. As the bellow figure illustrates, the data of repositories, merge scenarios, the commits in the scenario, merge replay with respects to different merging techniques, conflicting files and regions, code complexity, and code style violations are storied in our dataset.

Data Relations

Here we describe each of the tables separately.

Repository

Table Repository stores the information of repositories in the dataset. This information is retrieved with GitHub API.

Field	Data Type	Description
id	INT	The repository id on GitHub
update_date	DATETIME	The time that the information of the repository retrieved from GitHub and inserted in the dataset
name	CHAR(100)	The repository name
description	CHAR(400)	The repository description
url	ARCHAR(120)	The repository URL in <USER_NAME>/<REPOSITORY_NAME> format
language	VARCHAR(20)	The main programming language of the repository
watch_num	INT	The number of repository watches
star_num	INT	The number of repository stars
fork_num	INT	The number of repository forks
issue_num	INT	The number of repository issues
size	BIGINT	The size of the repository

Merge Scenario

The information of merge scenarios is stored in Merge_Scenario table.

Field	Data Type	Description
merge_commit_hash	CHAR(40)	The SHA-1 of the merge commit
ancestor_hash	CHAR(40)	The SHA-1 of the ancestor
parent1_hash	CHAR(40)	The SHA-1 of parent 1
parent2_hash	CHAR(40)	The SHA-1 of parent 2
parallel_changed_num	INT	The number of changes that edited in both parents
merge_commit_can_compile	INT	Whether the merge commit can compile;1 if can compile, 0 if cannot, and -1 if do not try to compile
merge_commit_can_pass_test	INT	Whether the merge commit can pass the tests;1 if can pass the tests, 0 if cannot, and -1 if do not try to run the tests
ancestor_can_compile	INT	Whether the ancestor can compile;1 if can compile, 0 if cannot, and -1 if do not try to compile
ancestor_can_pass_test	INT	Whether the ancestor can pass the tests;1 if can pass the tests, 0 if cannot, and -1 if do not try to run the tests
parent1_can_compile	INT	Whether parent 1 can compile; 1 if can compile, 0 if cannot, and -1 if do not try to compile
parent1_can_pass_test	INT	Whether parent 1 can pass the tests;1 if can pass the tests, 0 if cannot, and -1 if do not try to run the tests
parent2_can_compile	INT	Whether parent 2 can compile;1 if can compile, 0 if cannot, and -1 if do not try to compile
parent2_can_pass_test	INT	Whether parent 2 can pass the tests;1 if can pass the tests, 0 if cannot, and -1 if do not try to run the tests
merge_commit_date	DATETIME	The date of the merge commit
ancestor_date	DATETIME	The date of the ancestor
parent1_date	DATETIME	The date of merge parent 1
parent2_date	DATETIME	The date of merge parent 2
pull_request	INT	1 id the merge is a pull request, 0 otherwise

Merge_Related_Commits

The information of all commits in both parents from the ancestor to the merge commit is stored in Merge_Related_Commit.

Field	Data Type	Description
commit_hash	CHAR(40)	The SHA-1 of the commit
date	DATETIME	The date of the commit
message	VARCHAR(40)	The commit message
branch	VARCHAR(45)	The branch name of the commit
merge_commit_parent	INT	The parent that the commit is in
file_added_num	INT	The number of added files
file_removed_num	INT	The number of removed files
file_renamed_num	INT	The number of renamed files
file_copied_num	INT	The number of copied files
file_modified_num	INT	The number of modified files
line_added_num	INT	The number of added lines
line_removed_num	INT	The number of removed lines

Merge_Replay

We replay each merge scenarios to extract their characteristics. Since our dataset has the capability of storing the replying results with different merging techniques, we store the merge reply information in a separate table, called Merge_Replay.

Field	Data Type	Description
merge_technique	VARCHAR(15)	The merging techniques, the default is used for git merge with the default configurations
is_conflict	INT	1 if there is at least one conflict in the merge replay, 0 otherwise
can_compile	INT	Whether the replay can compile;1 if can compile, 0 if cannot, and -1 if do not try to compile
can_pass_test	INT	Whether the replay can pass the tests;1 if can pass the tests, 0 if cannot, and -1 if do not try to tun the tests
execution_time	FLOAT	The execution time of the merge replay in seconds
result_is_equal_to_replay	INT	1 id the merge replay is equal to the merge commit, 0 otherwise

Conflicting_File

For each merge replay, the data of files that have conflicts is stored in Conflicting_File table.

Field	Data Type	Description
file_path_name	VARCHAR(1000)	The relative path of the conflicting file
conflict_type	INT	1 if content conflict, 2 if rename conflict, and 3 if remove/edit conflict

Conflicting_Region

The data of conflicting regions is stored in Conflicting_Region table.

Field	Data Type	Description
parent1_start_line	INT	The start line of the conflict region in parent 1
parent1_length	INT	The length of the conflicting region in parent 1
parent2_start_line	INT	The start line of the conflict region in parent 2
parent2_length	INT	The length of the conflicting region in parent 2

Code-Style_Violation

The number of code violations is stored in Code-Style_Violation table.

Field	Data Type	Description
merge_commit_style_violation_num	INT	the number of code violations in the merge commit
ancestor_style_violation_num	INT	the number of code violations in the ancestor
parent1_style_violation_num	INT	the number of code violations in parent 1
parent2_style_violation_num	INT	the number of code violations in parent 2

Code_Complexity

The differences between the code complexity of two parents are stores in Code_Complexity table.

Field	Data Type	Description
merge_commit_hash	CHAR(40)	The SHA-1 of the merge commit
TODO_diff	INT	The differences of measure1 in two parents