As part of weekly cleanup activity we are identifying all the redshift tables which has tag as intermediate and truncating this from back end. But this does not refresh the sample dataset which users see while clicking on explore. Is there any specific cache/tmp folder can be cleaned up to make the explore dataset as blank after this routing activity.
the samples for the Explore tab are indeed cached on-disk, and since DSS has no fast and cheap way of checking whether a SQL table has been changed outside DSS, these cached samples are not invalidated.
The simplest to invalidate them is to make any small change on the dataset's settings and save (for example touching a comment on one of the columns of the dataset.