Hive enables MapReduce to be programmed using something that looks like SQL, instead of a procedural language like Java or Python. This is useful if a team of database, as opposed to application, programmers are called upon to program MapReduce.
Using Hive tables requires defining a schema.
The SQL-like language (called HiveQL) is converted to a MapReduce job.
Hue is a browser based GUI within which you can do Hive work. You type your query and see tabular results. Hue has ODBC drivers, and can export a CSV to Excel.
The Apache page for Hive calls it “a data warehouse system for Hadoop that facilitates easy data summarization, ad-hoc queries, and the analysis of large datasets.” I’m not sure how the data warehouse piece applies.