2.8 Human activities other than business processes 21 1. Mandatory column constraints: these require that a particular column in a table must have a value in every row (record) of that table: for example, each row in a table of purchase orders must include an order number and date raised as well as a customer number (a foreign key which implements a relationship to the table of customers); 2. Uniqueness constraints: these include the requirement for each row to have a unique identifier (e.g., an order number in the case of a table of purchase orders), otherwise records cannot be distinguished (we cannot rely on a combination of customer number and date raised, since the same customer may raise more than one order on the same day). These constraints also include those that are necessary properties of the real-world situation represented by the data: for example, every seat allocation record for a particular flight on a particular day must have a different seat number, otherwise there will be two passengers allocated to the same seat, neither comfortable nor safe! 3. Referential integrity constraints: these require that, in every row of a table, a particular column (or combination of columns) is either null (empty) or has a value or values that match the value(s) in a specific column or columns in either a. one of the rows of another table, or b. another row of the same table: for example, each customer number in a table of purchase orders must refer to an actual customer