Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2503.20593 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913759933497344 |
|---|---|
| author | Marcy, Mathilde Petit, Jean-Marc Scuturici, Vasile-Marian Bonjour, Jocelyn Fertel, Camille Cavalier, Gerald |
| author_facet | Marcy, Mathilde Petit, Jean-Marc Scuturici, Vasile-Marian Bonjour, Jocelyn Fertel, Camille Cavalier, Gerald |
| contents | Surrogate keys are now extensively utilized by database designers to implement keys in SQL tables. They are straightforward, easy to understand, and enable efficient access, despite lacking any real-world semantic meaning. In this context, complex redundancy issues might emerge and often go unnoticed as long as they do not affect the operational applications built on top of the databases. These issues become evident when organizations seek to leverage data science, posing significant challenges to the implementation of analytical projects.
This paper, grounded in real-world applications, defines the concept of artificial unicity and proposes RED2Hunt (RElational Databases REDundancy Hunting), a human-in-the-loop framework for identifying hidden redundancy and, if problems occur, cleaning relational databases implemented with surrogate keys. We first define the central and intricate notion of artificial unicity and then the RED2Hunt framework to address it. We rely on simple abstractions easy to visualize based on the so-called redundancy profile associated to some relations and the notion of attribute stability. Quite interestingly, those profiles can be computed very efficiently in quasi-linear time. We have devised different metrics to guide the domain expert and an actionable framework to generate new redundancy-free databases. The proposed framework was implemented on top of PostgreSQL. From the publicly available IMDB database, we have generated synthetic databases, implementing different redundancy scenarios, on which we tested RED2Hunt to study its scalability. RED2Hunt has also been tested on operational databases implemented with surrogate keys. Lessons learned from these real-life applications are discussed. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_20593 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | RED2Hunt: an Actionable Framework for Cleaning Operational Databases with Surrogate Keys Marcy, Mathilde Petit, Jean-Marc Scuturici, Vasile-Marian Bonjour, Jocelyn Fertel, Camille Cavalier, Gerald Databases Surrogate keys are now extensively utilized by database designers to implement keys in SQL tables. They are straightforward, easy to understand, and enable efficient access, despite lacking any real-world semantic meaning. In this context, complex redundancy issues might emerge and often go unnoticed as long as they do not affect the operational applications built on top of the databases. These issues become evident when organizations seek to leverage data science, posing significant challenges to the implementation of analytical projects. This paper, grounded in real-world applications, defines the concept of artificial unicity and proposes RED2Hunt (RElational Databases REDundancy Hunting), a human-in-the-loop framework for identifying hidden redundancy and, if problems occur, cleaning relational databases implemented with surrogate keys. We first define the central and intricate notion of artificial unicity and then the RED2Hunt framework to address it. We rely on simple abstractions easy to visualize based on the so-called redundancy profile associated to some relations and the notion of attribute stability. Quite interestingly, those profiles can be computed very efficiently in quasi-linear time. We have devised different metrics to guide the domain expert and an actionable framework to generate new redundancy-free databases. The proposed framework was implemented on top of PostgreSQL. From the publicly available IMDB database, we have generated synthetic databases, implementing different redundancy scenarios, on which we tested RED2Hunt to study its scalability. RED2Hunt has also been tested on operational databases implemented with surrogate keys. Lessons learned from these real-life applications are discussed. |
| title | RED2Hunt: an Actionable Framework for Cleaning Operational Databases with Surrogate Keys |
| topic | Databases |
| url | https://arxiv.org/abs/2503.20593 |