A record is a single entry in a database table. It contains all the fields for one entity. A customer record has a name, address, phone, and signup date. An order record has a customer ID, product, quantity, and price. The record is the unit of storage. When you insert a row, you create a record. When you query a table, you retrieve records. When you update a field, you change a record. The record is the atom of relational data.

Records relate to each other through keys. A customer record has a primary key. An order record has a foreign key pointing to that customer. The relationship links the two records without duplicating data. One customer can have many orders. Each order record references the customer. The customer record stores the customer's details once. This structure avoids redundancy and keeps data consistent. Records can also be nested in some databases. A document store might have a customer record with an array of order records inside it. That embeds the related data in one document. Embedded records are faster to read because they avoid joins. They are harder to update because the same data may appear in multiple documents. The choice between referencing and embedding depends on how the data is used. Relational databases reference. Document databases often embed. Both approaches have trade-offs.

Record characteristics

• Contains fields — one value per field
• Identified by a key — primary key or unique identifier
• Relates to other records — through foreign keys
• Stored as a row — in a relational table
• Updated atomically — changes commit or roll back together

A record is a fact about one thing. A table is a collection of related facts. A database is a collection of related tables.