We have a table name as sales, which is partitioned by period (YYYYMMDD), and we also need a table ly_sales(last year sales). To speed up the query, we don't use a view to join sales with last year mapping table( e.g 20110603 mapped to 20100603) for performance viewpoint. However we used the view which join the mapping table and then materialize the data into the table ly_sales. But since the data in hive is simple under some directory, such as
I am wondering whether we can create an empty table ly_sales, and the structure will be like
And then use soft link to link 20100603 under ly_sales to 20110603 under sales. Do you think it is doable?