tree 8078108f7c7a995cacbdf030e0c0836b63f6daeb
parent f8573ebedbeb3498a3cb5b5234131f913713d6ec
author Wes McKinney <wes.mckinney@twosigma.com> 1491463591 +0200
committer Uwe L. Korn <uwe@apache.org> 1491463591 +0200

PARQUET-946: Add ReadRowGroup and num_row_group methods to arrow::FileReader

There's a lot of room for improvement / further refactoring here -- the assumption that an entire column in a file is being read runs very deep in the Arrow reader, so I tried to do the minimum work to decouple the row group iteration. There's some code duplication in ReadRowGroup, but we should maybe save further cleanup for a future patch.

Author: Wes McKinney <wes.mckinney@twosigma.com>

Closes #291 from wesm/PARQUET-946 and squashes the following commits:

6d2b48a [Wes McKinney] Add virtual dtor
c7589f7 [Wes McKinney] Add ReadRowGroup and num_row_group methods to arrow::FileReader