What Is a Database?

A shoebox of receipts, a notebook of phone numbers and a stack of index cards all hold information. They also share a weakness. Finding one fact means flipping through everything, and nothing stops two entries from contradicting each other. A database is the fix for that weakness. It keeps information organized so that people and programs can look things up, add to it and change it quickly and reliably.

This entry explains what a database is, what software runs it, and where you already meet databases every day. A small library catalog serves as the running example. It does not teach the SQL language or compare database families in depth. Other entries in the Data and Databases area cover those topics.

A database is organized data with a manager

A database is an organized collection of data, stored electronically, that a program can search and update. Two ideas sit inside that sentence.

The first is organized. The data follows a known shape, so the computer knows where to look. In a relational database, which is the most common kind, that shape is the table: a grid of rows and columns. Each table is a named collection of rows, every row in a table has the same named columns, and each column holds one type of value, such as text, whole numbers or dates.

The second is the program. Raw data on a disk is not enough. A piece of software called a database management system, or DBMS, sits between the data and everyone who uses it. People often say "database" when they mean the whole package of data plus manager. The distinction is easy to remember with a building comparison: the books are the data, and the librarian who finds, lends and protects them is the DBMS.

A small example: a library catalog

Imagine a neighborhood library that wants to track its books. A table named books could look like this:

The titles and authors here are invented for the example. Each row describes one book. Each column describes one property of every book. The book_id column gives every row a unique label, so the library can refer to book 3 without confusion, even if two books share a title.

Now suppose a visitor asks for books by Mara Ellis that are on the shelf today. You can answer by eye with four rows. With forty thousand rows, you need a question the computer can run. Databases accept questions written in a language, most often SQL:

SELECT title
FROM books
WHERE author = 'Mara Ellis' AND on_shelf = 'yes'
ORDER BY title;

The result is:

The query describes what you want, not how to find it. The database decides the fastest way. That separation is one of the biggest reasons databases exist.

What the management software does

A DBMS does more than hold rows. Its main jobs are easy to list:

• Store. It writes data to disk in a form that survives a restart.
• Find. It answers questions quickly, even when the data is large.
• Change safely. It lets many people read and update the data at the same time without trampling each other's work.
• Enforce rules. You can require that a column is never empty, or that a value is not repeated. The database then refuses entries that break the rule.
• Protect. It controls who may read or change what, and it supports backups.

Why all-or-nothing matters

Some changes need several steps. If a library moves a book from the shelf to a borrower, it records the loan and marks the book as not on the shelf. If the system failed between the two steps, the records would disagree. Databases solve this with a transaction, a group of steps that either all happen or none happen. PostgreSQL's documentation describes this as atomicity and explains that, once a transaction is confirmed, its changes are safely recorded. You do not need to learn transactions now. You only need to know that this protection is built in, and that a plain file or a stack of cards cannot offer it.

Different kinds of databases

Tables are not the only shape. The main families, in simple terms:

• Relational databases store data in linked tables and are queried with SQL. PostgreSQL, MySQL and SQLite are well-known examples.
• Document databases store each record as a flexible, self-contained document, often written like JSON.
• Key-value stores work like a giant dictionary: you give a key and get a value back, which suits fast lookups.
• Graph databases focus on connections, such as who knows whom.

Treat the product names as examples. The ideas matter more than any brand. If you want to see when the relational and non-relational styles each fit, read relational vs NoSQL databases.

Databases also differ in how they run. Some run as a separate server that many computers connect to. Others, such as SQLite, are built into an application, and the whole database is an ordinary file next to the program. SQLite's own documentation lists embedded devices, desktop applications, data analysis and testing among its common uses.

Where you already use databases

You touch databases far more often than you notice. A bank balance, a seat reservation, a school grade, a product page with a stock count and a contact list on a phone all come from data kept in some database. When you search a shop for "blue jacket, size M", an application turns your clicks into a query and shows you the rows that match.

For anyone planning an IT path, databases appear in nearly every area: support teams look up accounts, developers store application data, analysts query records, and security teams search logs. Understanding the basics gives you a shared vocabulary for all of them.

Common mistakes

• Treating a database and a spreadsheet as the same thing. They overlap, but they are built for different jobs. Spreadsheet vs database explains the difference.
• Mixing up the data and the DBMS. A bug in the software and a mistake in the data are different problems with different fixes.
• Assuming a database is always correct. It enforces only the rules you give it. If a wrong value passes the rules, the database stores it faithfully.
• Thinking databases are only for large companies. A single-file database in a small app is still a database.
• Believing "database" means SQL only. SQL is the standard language for relational systems, but other kinds of databases use other query methods.
• Skipping backups. A database protects against many errors, but not against a deleted disk or a mistaken delete command. Keep copies.

FAQ

Is a database just a big file?

In the simplest case, yes, a database can live in one file. What makes it a database is the structure and the manager around it: rules, fast searching and safe updates.

Do I need to know SQL to use a database?

Not always. Many applications read and write the database for you. If you want to ask your own questions of the data, SQL is the common tool, and the entry on reading data with SELECT is a gentle start.

What is the difference between a table and a database?

A table holds one kind of thing, such as books. A database holds a set of tables and the links between them, such as books, members and loans.

Is Excel a database?

A spreadsheet can store data in rows and columns, but it does not enforce rules or handle many simultaneous editors the way a DBMS does. It is a useful tool for smaller jobs.

Which database should a beginner try first?

Any popular relational system works. SQLite needs no server and is easy to start with, while PostgreSQL is a common step up. The skills of designing tables and writing queries carry over between them.