You can write data to dynamic destinations (tables) - each table may contin separate schema version, eg: "table_v1", "table_v2", etc. Apache Beam or another procesing engine may be used. Next you can query the data with wildcard https://cloud.google.com/bigquery/docs/querying-wildcard-tables. "BigQuery uses the schema for the most recently created table that matches the wildcard as the schema for the wildcard table." - this could make the job, but you should ensure, that the table with the latest svhema version had been created last.