Reputation: 26759
http://wiki.apache.org/hadoop/Hive/HiveJDBCInterface - Is there a higher level ORM interface than the Hive JDBC driver which we can use to run the queries and parse the results
Upvotes: 4
Views: 9334
Reputation: 91
Hive is an eco-system component on top of Hadoop. It is not either a RDBMS or NOSQL Databases, it is just way of writing MapReduce programming on top of HDFS(filesystem).
Hive is created inorder to make developer friendly one who have sql knowledge can easily run MapReduce with just writing an sql kind of statements. The MapReduce can be expressed as an sql where Hive engine convert them to MapReduce using Driver.
Hive make division of storage and its structure using metastore. Metastore is simple RDBMS, like any other RDBMS we can connect to it and make an sql statement on top of this.Once we made an HQL statement on Metastore, the metastore have a capability to tell underline Hive engine to execute HQL statement and gives back the result as a resultset.
Upvotes: 0
Reputation: 6289
No. The "R" in ORM stands for a "relational database". Hive is a NOSQL databse, not a relational database. It's a huge advantage of Hive comparing to many other NOSQL databases that you can access it using (almost) normal SQL and using (almost) full featured standard JDBC interface. But if you are thinking about using ORM and Hive together, you might be approaching your problem from an incorrect angle.
EDIT 6/4/13: Here's a quote from Wikipedia: In the relational model, each table schema must identify a primary column used for identifying a row called the primary key. Tables can relate by using a foreign key that points to the primary key.
Hive does not support neither foreign nor primary keys, so it is not a relational database.
I would like to thank the downvoter for giving me an opportunity to doublecheck the reasoning that led to making my original statement.
Upvotes: 8