By default, a simple query in Hive scans the whole Hive table. This slows down the performance when querying a large-size table. The issue could be resolved by creating Hive partitions, which is very similar to what's in the RDBMS. In Hive, each partition corresponds to a predefined partition column(s) and stores it as a subdirectory in the table's directory in HDFS. When the table gets queried, only the required partitions (directory) of data in the table are queried, so the I/O and time of query is greatly reduced. It is very easy to implement Hive partitions when the table is created and check the partitions created, as follows:
-- Create partitions when creating tables jdbc:hive2://> CREATE TABLE employee_partitioned . . . . . . .> ( . . . . . . .> name string, . . . . . . .> work_place ARRAY<string>, . . . . . . .> sex_age STRUCT<sex:string,age:int>, . . . . . . .> skills_score MAP<string,int>, . . . . . . .> depart_title...
