Data becomes easier to validate, query and exchange when repeated pieces of information follow a known structure.

A record groups related fields. For example,

name: Ada
age: 36
active: true

contains three fields with different meanings and value types.

Schemas

A schema describes the expected structure of data: which fields exist, what kinds of values they hold and which constraints apply.

For example, a user record might require a unique identifier and a name while allowing an optional email address.

Structure and serialization are different

The same logical record can be serialized in different formats such as JSON, CSV rows or database columns.

A serialization format determines how data is represented as text or bytes. A schema determines what the represented values mean and how they fit together.

Validation

Incoming data can be checked against its schema before the rest of a program relies on it. Validation can reject missing required fields, values of the wrong type or values outside permitted ranges.

Collections of records

When many records share a schema, software can process them systematically: select fields, compare records, group them and store them in tables or other structured collections.

Schemas turn informal data conventions into explicit structure that other systems can depend on.