The Data Hierarchy Attributes, Records, and Files
Attributes of Accounts Receivable Customer Account Number (Key) Customer Name Customer Address Current Balance of Account Customer Credit Limit Accounts
Receivable Record:
All Accounts Receivable Records
Accounts 1: Receivable Record Accounts Receivable File relevant because it contributes to the information content of the entire set. As proof of this, the absence of any single relevant attribute diminishes or destroys the information content of the set. The addition of irrelevant or illogical data would not enhance the information content of the set.
Record. A record is a complete set of attributes for a single occurrence within an entity class. For example, a particular customer’s name, address, and account balance is one occurrence (or record) within the AR class. To find a particular record within the database, we must be able to identify it uniquely. Therefore, every record in the database must be unique in at least one attribute.1 This unique identifier attribute is the primary key. Because no natural attribute (such as customer name) can guarantee uniqueness, we typically assign artificial keys to records. The key for the AR records in Figure 1-6 is the customer account number. This is the only unique identifier in this record class. The other attributes possess values that may also exist in other records. For instance, multiple customers may have the same name, sales amounts, credit limits, and balances. Using any one of these as a key to find a record in a large database would be a difficult task. These nonunique attributes are, however, often used as secondary keys for categorizing data. For example, the account balance attribute can be used to prepare a list of customers with balances greater than $10,000.
Files. A file is a complete set of records of an identical class. For example, all the AR records of the organization constitute the AR file. Similarly, files are constructed for other classes of records such as inventory, accounts payable, and payroll. The organization’s database is the entire collection of such files.
When we get into more advanced topics, we will see how a combination of nonunique attributes can be used as a unique identifier.