Recently myself and Mitchell Pearson (blog|twitter) were working on a project for a client that required us to load a ton of data (dozens of TBs) into some tables each built with a clustered columnstore index. We discovered during testing that the fastest way to get that much data into the clustered columnstore index is to create an empty uncompressed table, load the data into the uncompressed table, then apply the clustered columnstore index to the table, and partition switch the data into the main table. In order to facilitate this, I created this script to dynamically create a copy of the target table (without the columnstore index), create the clustered columnstore index, and then do the partition switch automatically.