The main aim of this work is to create a PostgreSQL extension, which can be installed to a database by CREATE EXTENSION statement, for uniforming the names of base table constraints. First, I conducted a research, based on the PostgreSQL databases of three open source systems (FusionForge (version 6.1beta1) [1], LedgerSMB (version 1.5.20) [2] and OTRS (version 6.0.6) [3]), to find out used conventions for naming constraints and what is the state of the art of naming database objects in general.
Good intention-revealing names and following certain naming conventions are a part of clean code [4] practices. Naming things is one of the hardest things in Computer Science [5]. Bad naming habits make administrating and further development of any system, including databases, significantly more difficult and it also makes it harder to read and understand the code. Specifically, bad naming of database constraints makes the administration of the database schema more complex and also makes it tricky to understand error messages as well as execution plans of data manipulation statements.
From the analysis, it turned out that none of the observed databases had complete consistency in naming constraints. There were many different patterns by which the constraints were named. That kind of inconsistency makes database administration more difficult. Therefore, an extension for uniforming constraints names would certainly be a useful tool for any database developer.
The thesis is in Estonian and contains 52 pages of text, 6 chapters, 19 figures, 8 tables.
[1] FusionForge [WWW] https://fusionforge.org/ (18.04.2018)
[2] LedgerSMB [WWW] https://ledgersmb.org/ (18.04.2018)
[3] OTRS [WWW] https://otrs.com/ (18.04.2018)
[4] Martin, R.C. Clean Code. A Handbook of Agile Software Craftsmanship. Pearson Education, 2009
[5] Fowler, M. TwoHardThings, 2009 [WWW] https://martinfowler.com/bliki/TwoHardThings.html (19.04.2018)