I\'m creating a dataflow pipeline. To minimize query cost, it is supposed to read a very big table in BigQuery (once), then write it into two destination tables in Cassandra
