Redshift will ALWAYS load data into unsorted region after the initial table population, except in one very specific scenario. Load all your data in a single initial COPY command and it will come in all sorted, don't run one command per file or folder.