Migrate bulk-export code to use DataFusion query code path instead of compaction code path.

[m09526]

When the bulk export code path was created, the DataFusion code path only existed for compactions not queries so it uses that code path directly. As this is going to be a query side enhancement it makes more semantic sense for this code to use the DataFusion query code paths. That code still includes options to output results to a Parquet file(s) rather than Arrow RecordBatches back to the client.