Rowtype

Rowtype is a data concept used in several relational database systems to represent a single row of a table or query result as a single, composite value. It provides a convenient way to work with an entire row without declaring separate variables for each column. In practice, rowtype is used to declare variables that can hold all the fields of a row and to fetch or manipulate whole rows at once.

In Oracle, the term is typically expressed as the %ROWTYPE attribute. A variable declared with a table’s rowtype mirrors every column of that table, including its data types. For example, a variable declared as employees%ROWTYPE will have fields corresponding to all columns of the EMPLOYEES table. This is useful for fetching a row into a single variable or iterating over rows returned by a cursor. Access to fields is by dot notation, such as v_emp.id or v_emp.salary. If the table schema changes, you generally need to recompile the PL/SQL code to reflect the new structure.

In PostgreSQL, a similar concept exists in PL/pgSQL as table%ROWTYPE. Declaring a variable like rec employees%ROWTYPE creates a composite value whose fields match the columns of the table. You can populate it with a SELECT ... INTO rec, or fetch into it from a cursor, and then access fields as rec.id, rec.salary, etc. The advantage is automatic alignment of the variable’s fields with the table’s column types. Changes to the table may require recompilation of the function to reflect new or removed columns.

Rowtype offers type-safe, convenient handling of entire rows, at the cost of coupling to a specific table’s schema. It is most valuable when operations naturally process full rows rather than individual columns. See also: ROW/RECORD types, table expansion considerations, and schema evolution impact.