Bridging legacy ODBC drivers to Apache Arrow memory structures opens up huge zero-copy data transfer pipelines for analytical workloads. Great implementation in C for minimal overhead.