refactor(odbc): implement bulk fetch using columnar buffers

This commit introduces a new bulk fetch implementation in the ODBC bridge, utilizing columnar buffers for improved performance. The changes include the addition of a `build_bindings` function to create column bindings, a new `stream_rows` function to handle batch processing, and the extraction of values from buffers. This refactor enhances the efficiency of data retrieval by reducing the overhead of row-by-row fetching.

Author: lovasoa