I want to create a Apache Beam Dataflow Pipeline which will read the parquet files from GCS bucket and load into BQ. I can specify the Dataset and table name but cannot spec