adjust ID of duplicate NA patterns

[benthestatistician]

One of the tasks of design_matrix() (aka model_matrix()) is to identify terms with common missingness patterns, so as to avoid storing the same information in two places. Some of the operations would be more succinct if they used duplicated.matrix(..., margin=2) to identify replicate columns of the data frame that is cast but not stored here --

null.record <- rowSums(as.data.frame(ccs.by.term))==0

-- rather than applying duplicated.default() to a list. They may well also be more efficient; the duplicated() help page bears a "Warning" reading:

Using this for lists is potentially slow, especially if the
elements are not atomic vectors (see ‘vector’) or differ only in
their attributes. In the worst case it is O(n^2).