1151 commits with software maintenance activity labels (corrective,perfective,adaptive)

Data format: CSV

Separator character: '#'

This dataset contains 1151 commits manually labeled with maintenance activities ("c" for corrective, "p" for perfective, "a" for adaptive) according to the definition by Mockus et al. in  "Mockus, A. and Votta, L.G., 2000, October. Identifying Reasons for Software Changes using Historic Databases. In icsm (pp. 120-130)".

In addition, this dataset also contains further information (features) extracted from the commits:

The source code changes performed by the commit author as part of a given commit (statement added, statement removed, etc.)

The source code change taxonomy is detailed in "Fluri, B. and Gall, H.C., 2006, June. Classifying change types for qualifying change couplings. In Program Comprehension, 2006. ICPC 2006. 14th IEEE International Conference on (pp. 35-45). IEEE."

A binary indication (1/0) whether a given commit contains any of the keywords from a pre-computed (according to a word frequency analysis) set of keywords indicative of each maintenance activity.

The dataset consists of commits sampled from the following open source projects:

RxJava

hbase

elasticsearch

intellij-community

hadoop

drools

kotlin

restlet-framework-java

orientdb

camel

spring-framework    

This dataset is a supporting material for the paper "Boosting Automatic Commit Classification Into Maintenance Activities By Utilizing Source Code Changes", to appear in PROMISE 2017.