What are projection operations?

Projection is the operation of converting an object into a new form, usually containing only those attributes that will be used later.

By using projection, you can construct new types that are generated on a per-object basis. You can map an attribute and perform mathematical functions on that attribute. You can also map an object without changing the original object. Simply put, this means selecting certain attribute columns in a relationship.

Projection of a relation R is the process of selecting a number of attribute columns from the relation R to form a new relation in two steps:

(1) Select the specified attributes to form a table that may contain duplicate rows.

(2) Delete the duplicate rows to form a new relationship.

Extension

The relational schema R(A1,A2,.... ...An), t∈R denotes that t is a tuple of R, where A is a column of attributes in R. t[Ai] then denotes a component of tuple t corresponding to attribute Ai. The arithmetic expression for projection is: ?

Projection in a relational two-dimensional table is a vertical operation that targets the attribute columns in the two-dimensional table.

In SQL, the projection operation is a unary operation that selects the desired columns from a relation R to form a new relation

Ref:

Baidu Encyclopedia - Projection Operations