This article introduces the concept of the Hive.
Summary:
Chinese called Hive data warehouse management system, before we MapReduce operation must either be achieved through special commands by writing code, we will be able to operate with Hive MapReduce cluster by commonly used SQL statements. I am not feeling very convenient. It is also easy to understand MapReduce principle, people understand SQL statements used.
There are several companies have launched their own Hive, the more popular is Apache Hive , CDH Hive, HDP Hive and MapR Hive, we have just started to learn Apache Hive mostly used, but companies rarely use it because it's too messy version, which is also a lot of BUG, can not be quickly put into production, so most are using third-party Hive, which is CDH or MapR Hive, the Hive, developed by specialized organizations, conditioning clear, less BUG, of course, people this service is also relying on friends to make money. Bloggers also because the learning stage, first introduced Apache Hive, and follow-up will introduce and set up other versions.
structure:
Setp1: the user through the Shell command, WebUI or call JDBC Driver
Setp2: Driver will go to the database query has no information on this table, if not directly returned, yes, the third step
Setp3: turn the SQL command execution behavior MapReduce
Setp4: to distribute to execute MapReduce