A primary key that serves no other purpose.

In the case a relational database table contains no single “best” Primary Key (or even no single possible Primary Key), a new field is added whose only purpose is to uniquely identify a row. These fields are often named “ID” or something to that effect. Fields whose only purpose in a table is to allow each row to be uniquely identified are called “Surrogate Keys”.

Surrogate keys are necessary to prevent a collisions stemming from coincidence. For example: you have a student ID because your name is Joshua Brown and there’s another “Joshua Brown” in your class.

Surrogate keys may often just be more convenient. You could concoct a crazy combination of fields that could, in aggregate, be called a primary key… but it’s just easier to add an “ID” column and call it good.

Examples

  • Social Security Numbers
  • Student # or Employee #
  • Account numbers

Source

Sources

Primary key