Open njtierney opened 6 years ago
And implementing this for cases, rowise, would resolve #248
A big architectural question you may have already thought about: what if you had a utility function as_naniar(df, as_missing = c(NA))
that returned a new class that included the tibble but also defined which values to count as missing? Then, users wouldn't need to specify missing values each time they called a function (as is currently implemented) and could still specify different missing values for each dataset (which couldn't happen if you used an options approach at the start of each session). Would still inherit tbl
class for interoperability with other tidy packages.
the current output of miss_scan_search is:
Created on 2018-06-28 by the reprex package (v0.2.0).**
But I think that it should be different, something like this:
Would be more useful. This becomes more apparent if you had something like this:
And this:
And possible a summary on the bottom: