fct_lump all with count above/below threshold

[RoyalTS]

It would be great if fct_lump() had an additional mode that lumped factor levels not based on the final number of factor levels desired (n) or the proportion of the data that accrues to the level (prop) but based on the raw count for the level (call it count?).

Stealing the example data from #43:

N <- 5
M <- 2 ** N
exp_factor <- factor(rep(1:(N + M), c(2 ** (N:1), rep(1, M))))
table(exp_factor)
## exp_factor
##  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 
## 32 16  8  4  2  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1 
## 26 27 28 29 30 31 32 33 34 35 36 37 
##  1  1  1  1  1  1  1  1  1  1  1  1

Desired behavior: forcats::fct_lump(exp_factor, count = 1) would retain only the factor levels 1:5 and Other.

[AshesITR]

That's a good idea. Maybe naming the argument min_count with comparison >= makes the functionality clearer...

[hadley]

I think this is a reasonable request, but my main concern is making fct_lump() too complicated. It might be time to start spinning out fct_lump_prop() etc.

[robinsones]

👋 I'll try it