Hive

• Allows fetching data from HDFS using SQL queries
• Hive can be used with MapReduce or Tez to fetch the data
• It's a translation layer from SQL to MapReduce/Tez commands
• Uses HiveSQL

Drawbacks

• Simple OLAP queries! (read operations). Bad for OLTP (write operations)
• Stores data de-normalized
• Limited to what SQL can do (pig, spark allows more complex stuff)
• No transactions

Using Hive

• Hive CLI
• Saved query files (hive -f /home/user/queries.hql)
• Ambari / Hue
• JDBC/ODBC server
• Thrift service (even though hive is not suitable for OLTP)
• Via Oozie