From the hive perspective, the following are the considerations:
1. From which nodes do you plan on running the queries? If it is from one of the nodes and you plan on running queries from a particular directory all the time, installing the hive tarball/rpm only on that node suffices. Although this is not very practical (one node same directory all the time), it is a quick and dirty solution. If you plan to use ambari (which is what we recommend), you should be able to easily configure and start hive.
2. If you plan on executing queries from more than one node and or directory, you need to setup a mysql server. This can also be setup by ambari if you choose to go that route.
If you are going the rpm route, this link should give you more information on how to get things up and running:
Pay particular attention to the Set up hive/hcatalog configuration files section.
Coming to your 2nd question, if you mean to load data that is already in HDFS to hive, you can do that via the LOAD command on the hive command line.
Please let me know if that answers your questions.