Exploring views

Views in SQL provide abstraction from querying a table directly. A view could be a combination of multiple tables joined or grouped on a set of columns. In RDBMS, views could be broadly categorized into two types:

• Materialized views

• Non-materialized views

Hive supports only non-materialized views and as it does not support materialized data, the view is strictly bound to the tables it is based on. In other words, it also means if the columns of the tables are altered or dropped, it would affect the view or even fail the view.

How to do it…

These views are the logical constructs that do not store data with them. When a view is created in Hive, the underlying query is stored in the metastore. When the view is queried, the view's clauses or conditions are evaluated before the underlying query clause. For example, if the query has a limit of 200 and the view has a limit of 100 then the query would return 100 results.

An example of a view definition is as follows:

CREATE VIEW sales_view...