bigquery: Accelerate query result reading with the BigQuery Storage Read API #7031

Accelerate RowIterator for large query result sets by automatically streaming results in Apache Arrow format over gRPC via the BigQuery Storage Read API (client::Read), while keeping the user-facing RowIterator and Row API unchanged.

Currently, RowIterator fetches subsequent pages of query results sequentially via BigQuery v2 with JSON pagination (jobs.getQueryResults), which becomes a bottleneck on medium-to-large result sets.

• bigquery: Expose Arrow C Stream interface (FFI_ArrowArrayStream) on CompleteQuery
  • To be discussed if we should do it now