Given a tibble x of variables, some of which are wide form category counts, and the remainder of which are document variables (docvars in quanteda terminology), decompose the variables specified by the tidyselect arguments into a 'counts' variable in sparse format and keep the remainder.
jl_process_topics(x, ..., drop_old_vars = TRUE, drop_unused_vars = TRUE)
| x | a tibble |
|---|---|
| ... | a tidy select expression |
| drop_old_vars | whther to remove the wide variables (default: TRUE) |
| drop_unused_vars | whether to drop wide variables whose column sums are zero |
a tibble with a 'counts' variable