Exclusion filters express logic that defines a set of rows that should be dropped from the dataset. The rows aren't permanently deleted---you can recover them at any time by removing the exclusion filter---but they are omitted from all views and calculations, as if they had been deleted.
exclusion(x)
exclusion(x) <- valuea Dataset
an object of class CrunchLogicalExpr, or NULL
exclusion returns a CrunchFilter if there is one,
else NULL. The setter returns x with the filter set.
Note that exclusion filters work opposite from how "normal" filters work. That is, a regular filter expression defines the subset of rows to operate on: it says "keep these rows." An exclusion filter defines which rows to omit. Applying a filter expression as a query filter will have the opposite effect if applied as an exclusion. Indeed, applying it as both query filter and exclusion at the same time will result in 0 rows.