Why Databases Exist

Databases were not invented because someone thought they were elegant. They were invented because, without them, working with data at any real scale becomes a nightmare. Let us look at the specific nightmares.

Problem 1: Keeping data without losing it

Imagine a shop that records every sale in a notebook. The notebook can be lost, burned, soaked, or simply run out of pages. If two clerks each keep their own notebook, the shop's records are now split across two places that may disagree.

A database provides durable, central storage: one trusted place where data lives, survives crashes, and can be backed up.

Problem 2: Many people at once

Now imagine ten clerks all trying to update the same inventory count at the same moment. Two of them sell the last umbrella within the same second. Who got it? Without coordination, the data ends up wrong — the system might sell the same umbrella twice.

Databases solve this with concurrency control: they let many users read and write at the same time while keeping the data consistent, so the last umbrella is sold exactly once.

Problem 3: Finding things quickly

A notebook with 100 sales is easy to scan. A notebook with 10 million sales is not. Answering "how much did we sell in March to customers in Ohio?" by reading every line by hand would take days.

Databases are built to answer questions fast, even over huge amounts of data, using clever internal structures (we will meet the idea of an index later) so that finding the right rows does not require reading every row.

Problem 4: Keeping data consistent

Suppose a customer's address is written in five different places. A customer moves. Now you must remember to update all five — and if you miss one, your data quietly contradicts itself. Which address is correct? Nobody knows.

Databases let you store each fact once and refer to it from everywhere else, so updating it in one place updates it everywhere. This idea — don't repeat the same fact — is at the very heart of relational databases, and we will return to it many times.

Problem 5: Protecting the data

Data must be protected from two kinds of harm: accidents (a program crashes halfway through a sale) and bad values (someone types a price of "banana" or an age of -7).

Databases protect against both. They can guarantee that a change either fully happens or does not happen at all (no half-finished sales), and they can enforce rules about what counts as valid data, refusing values that break the rules.

The big picture

Every feature you will learn in this course exists to solve one of these five problems. When a concept feels abstract, ask yourself: which of these nightmares is this protecting me from? The answer is almost always illuminating.

Check your understanding

QuestionSelect one

Storing a customer's address in five different places creates which classic problem that databases are designed to prevent?

The data takes up too much disk space to store.

Inconsistency — if the address changes, some copies may be updated and others missed, so the data contradicts itself.

The address becomes impossible to search.

The database runs out of memory.

QuestionSelect one

Two clerks try to sell the last item in stock at the same instant. Which database capability prevents the item from being sold twice?

Fast searching over large tables.

Backing up data to survive crashes.

Concurrency control — coordinating simultaneous reads and writes so the data stays consistent.

Enforcing that prices are numbers, not text.