How do you design a database when the tables start adding up?

November 17, 2026 5 answers
Asked by 2q+claynotch_817

I'm planning a database with at least 15 tables, and keeping track of relationships, indexes, and future features is getting tricky. How do you organize the design without trying to predict everything upfront?

Answers

Varnix-9e

Start with the data and the questions the application needs to answer. For example, which orders belong to a user? Write down a handful of those questions, then sketch the entities and their relationships on paper.

Mark which relationships are one to many and which are many to many. You don't have to keep the whole schema in your head, and a drawing is much easier to revise than a pile of tables.

u.479~draincaper:

It can help to keep the sketch focused on one part of the application at a time, then connect those pieces as the relationships become clearer.

Reply to this answer
_v95~canvasbadge.4

Learn the basics of normalization, including third normal form. A useful practical question is which table owns each piece of information. If the same fact is stored in multiple places, decide which copy is authoritative and how the others stay current.

Start with a normalized design and consider denormalizing when you have a measured performance reason. It can make reads simpler or faster, but duplicated data also creates more work when something changes.

Reply to this answer
3u~bracketkettle_672

Don't add an index to every column just because it might be queried someday. Start with the queries you know you need, then check how they perform as the data grows. Indexes help certain reads, but they also have storage and write costs.

Relationships and foreign key constraints are a separate decision from indexes. Some systems leave relationships unenforced in the database, but that means the application or other processes need to prevent inconsistent references.

Reply to this answer
4r!pixelratchet_815

Fifteen tables isn't inherently a large database. Table count alone doesn't tell you how much data you'll have or which queries will be expensive.

Design for the requirements you understand now and leave room to revise the schema. Trying to map every possible future feature before building anything can make the job much harder than it needs to be.

Reply to this answer
_d73~orvani.4

For reading, look into database normalization and database architecture. Designing Data-Intensive Applications is another book to consider if you want a broader look at how data systems handle growth.

For a less settled part of the model, JSON is an option, but it doesn't remove the need to decide what the data means. You can refine that area later as the requirements become clearer.

Reply to this answer

Add your answer

Answers are reviewed before publication.

Related Discussions