The database is at the top level: it is a standalone entity encompassing all data that may be in connection to each other by any means.

Database tables consist of a header and data rows, and each row has exactly identical inner structure. Data rows are also called records in database terminology.

This figure contains a small part of the Nuclide table. Several data records are contained, and each row describes a unique Nuclide entity. Records may be queried, updated, inserted or deleted in a data table.

A record comprises of several fields. All records in a table must have fields of identical name and data type. For example, a Nuclide record contains the following data fields:

Database management systems are required to retrieve, insert and delete records efficiently. To accomplish this task, each record must be unique, that is, must be identified unequivocally within a table. For identification purposes, one or more field values are used together, thus forming an identifier, or primary key.

For example, unique identifier of the nuclide table is a special field named NUCLIDE_ID. As it is declared unique, the database management system guarantees that duplicated NUCLIDE_ID will never be kept in the Nuclide table.

Please note that the A, Z field combination also would have been used as a primary key, but a single primary key was chosen because of its simpler usage.