In the documentation for the slice_min() and slice_max() functions it says that the order_by argument can be a variable or function of variables to order by.
What is meant by function of variables, and how could this be applied in a practical sense? Could it, for example, be used to supply a custom order of categorical values?
I've tried what feels like an exhaustive search for info online, but to no avail, so I'm turning to you good folk. Thank you.
What is meant by function of variables, and how could this be applied in a practical sense?What is meant by function of variables, and how could this be applied in a practical sense?
I think the most common usage of "a function of variables" would be any function that you give as input columns from the data frame, and it returns a numeric result (or at least something that has a "max" value). Here are a couple examples:
## get the row with the highest product of Sepal.Length and Sepal.Width
iris %>% slice_max(Sepal.Length * Sepal.Width)
## here we use the function `*` and the variables `Sepal.Length` and `Sepal.Width`
iris %>% slice_max(nchar(Species))
## get the rows with the longest species name
## here we use the function `nchar` and the variable `Species`
Could it, for example, be used to supply a custom order of categorical values?
Generally, if you want a custom order for categorical variables, we use factor and specify the order of the levels. Yes, you can use this within slice_max - the last factor level is considered the max:
iris %>% slice_max(Species)
## defaults to alphabetical order - all virginica rows returned
iris %>% slice_max(factor(Species, levels = c("versicolor", "virginica", "setosa")))
## if we make "setosa" the last/max level than setosa rows will be returned