How is Life for a Table in an Evolving Relational Schema? Birth, Death & Everything in Between

Panos Vassiliadis, Apostolos Zarras, Ioannis Skoulis

34th International Conference on Conceptual Modeling (ER 2015). 19-22 October 2015, Stockholm, Sweden

Summary

As a database evolves over time, its schema evolution affects all its surrounding applications, both semantically and syntactically. Studying the patterns and commonalities governing the evolution of database schemata is of great importance, as it can facilitate the proactive response to expected events and reduce maintenance effort and costs. Due to the non-availability of schema histories, till now, the related literature contains few small-scale studies of schema evolution. We have exploited the availability of the version history of databases that are part of larger open source projects, publicly available through open source repositories, and, we have performed a thorough, large-scale study on the evolution of eight such databases. We report on our observations and patterns on how evolution related properties, like the possibility of deletion, or the amount of updates that a table undergoes, are related to observable table properties like the number of attributes or the time of birth of a table.
Our findings indicate that (i) most tables live quiet lives; (ii) few top-changers adhere to a profile of long duration, early birth, medium schema size at birth; (iii) tables with large schemata or long duration are quite unlikely to be removed, and, (iv) early periods of the database life demonstrate a higher level of evolutionary activity compared to later ones.

Plz., refer to our Schema biographies page for a general overview of our research program.

Texts

Panos Vassiliadis, Apostolos Zarras, Ioannis Skoulis. How is Life for a Table in an Evolving Relational Schema? Birth, Death & Everything in Between. 34th International Conference on Conceptual Modeling (ER 2015). 19-22 October 2015, Stockholm, Sweden

[Local copy of the paper at ER 2015 (PDF)]

Panos Vassiliadis , Apostolos V. Zarras , Ioannis Skoulis. Gravitating to Rigidity: Patterns of Schema Evolution -and its Absence- in the Lives of Tables. Information Systems, Volume 63, January 2017, Pages 24-46, ISSN 0306-4379, doi:10.1016/j.is.2016.06.010.

Long version of the ER 2015 paper, [Local copy of the paper IS 2017 paper (PDF)]

Presentations

Experimental Resources

The following code and data are presented on-line to allow the reproduction of results by others. We would like to to clearly state that we simply cannot support any requests for the maintenance of the code, or clarifications, explanations etc. Moreover, we do not assume any responsibility for any side effects of the code (although we cannot think of, or have ever encountered, any). You are free to reuse the following code and data for academic purposes, provided you give the appropriate citation:

Panos Vassiliadis, Apostolos Zarras, Ioannis Skoulis. How is Life for a Table in an Evolving Relational Schema? Birth, Death & Everything in Between. 34th International Conference on Conceptual Modeling (ER 2015). 19-22 October 2015, Stockholm, Sweden. Source code, datasets, presentations available at http://www.cs.uoi.gr/~pvassil/publications/2015_ER/

(and, yes, academic honesty rules impose that this includes student projects too ;) )

Input: Database history versions (Raw input for extracting transitions)

Code: Source code for Hecate. Requires Java 7 and Eclipse.

Results: Spreadsheets for the eight data sets