Nezih Yigitbasi

Presto also handles case sensitivity itself on the task side, please see [this](https://github.com/facebook/presto/pull/4686) PR.

@spena thanks for taking the time to create this PR. Can you also update the README of the benchmark module?

@spena Added a few comments. Once you update the README and address the comments, I will be happy to give it a try. Thanks again!

@spena To keep the fixes to the existing benchmark module separate I created a new PR: https://github.com/apache/parquet-mr/pull/226 Yep I verified that your change makes the read-benchmark go through.

@spena I only added a few minor comments to the README. Other than that LGTM. @rdblue I have provided feedback to @spena and verified his changes. When he addresses my...

@davies in my microbenchmarks I saw ~2x improvement for primitive types (since for primitives the current implementation bypasses the converters) while the performance was roughly the same for complex types....

@julienledem Thanks for the comments, I have made a first pass over them. Regarding reducing the responsibility of `RecordReader`, I moved the vector-related interface to a new interface (`VectorizedRecordReader`, see...

@julienledem I have addressed most of your comments and fixed the unit tests. Can you take another look?