Querying pail data files using Mapreduce

[pankug]

I am having a problem in Lambda Architecture, Our data stored in HDFS is in fact based pail format using Thrift Serialization schemes and vertical partitioning.

Is there any direct way we can query our data (residing in HDFS) in batch view, so we don't have to store our output data in ElephantDB or another database and we can directly view our data and stored it in a readable format.

Is there any example for querying using map-reduce , PIG or HIve without using cascalog.

We are facing this problem If u can help us regarding code structure and other techniques using any set of Big Data tool and language