What are the differences between primary key foreign key and candidate key?

A primary key is a column which uniquely identifies the records in a table. In a broad sense, a primary key is the mixture of a unique key and an index: A column with a primary key is indexed to deliver a faster query, and doesn't allow duplicate values to ensure specific data. Most programmers recommend all tables having a primary key (and only one) to enhance the speed of queries and overall database performance. An example of a primary key may be found in a table named "departments," which might have a column named "department_number" that uniquely identifies each department in the table with a number.

A foreign key is a column (the child column ) in a table which has a corresponding relationship and a dependency on another column (the parent column ) that is usually in a different table. Parent columns can have multiple child columns, but a child column can only have one parent column. The child column is the column with the foreign key; the parent column does not have the foreign key "set" on it, but most databases require the parent column to be indexed. Foreign keys are made to link data across multiple tables. A child column cannot have a record that its parent column does not have. Say a table named "employees" has 20 employees (rows) in it. There are 4 departments in the "departments" table. All 20 employees must belong to a department, so a column in the "employees" table named "department" would point to the primary key in the "departments" table using a foreign key. Now all employees must belong to a department as specified by the "departments" table. If a department isn't specified in the "departments" table, the employee cannot be assigned to it.

A candidate key would be any key which could be used as the primary key, which means that the combination of the columns, or just the single column would create a unique key. You would then need to determine which of these candidate keys would work best as your primary key.