In Hive, views are logical data structures that can be used to simplify queries by either hiding the complexities such as joins, subqueries, and filters or by flatting the data. Unlike some RDBMS, Hive views do not store data or get materialized. Once the Hive view is created, its schema is frozen immediately. Subsequent changes to the underlying tables (for example, adding a column) will not be reflected in the view's schema. If an underlying table is dropped or changed, subsequent attempts to query the invalid view will fail, as follows:
jdbc:hive2://> CREATE VIEW employee_skills . . . . . . .> AS . . . . . . .> SELECT name, skills_score['DB'] AS DB, . . . . . . .> skills_score['Perl'] AS Perl, . . . . . . .> skills_score['Python'] AS Python, . . . . . . .> skills_score['Sales'] as Sales, . . . . . . .> skills_score['HR'] as HR . . . . . . .> FROM employee; No rows affected (0.253 seconds)
When creating views, there is no MapReduce job triggered at...
