Reduce Redshift tables created by this code

[sandramchung]

This code currently creates a new table in our Redshift cluster for every survey, which is a problem since the cluster is limited to 9900 tables and we hit that limit and have to delete lots of tables every few months. There are currently more than 1300 tables in our cluster created by this code. Please modify this code to save all survey results to a single table keyed by survey ID and/or periodically drop old survey results tables.

[sandramchung]

Per discussion, this should pull survey results from only surveys that have gotten results in the last 2 months, and drop the rest.