Normalizes numerical values into [0, 1]. For a data frame, matrix,
or tibble, each numeric column is normalized independently against its
own min/max, and non-numeric columns are dropped — matching the
original implementation this was migrated from.
Value
A normalized object: a numeric vector stays a vector; a matrix stays a matrix; a data frame or tibble is reduced to just its (now normalized) numeric columns.
Details
Missing values are preserved and ignored when finding the min/max;
only finite values define the range. A constant input (zero range) maps
every finite value to 0 instead of NaN, and an input with no
finite values comes back as all NA.
Examples
min_max_norm(c(10, 20, 30, 40, 50))
#> [1] 0.00 0.25 0.50 0.75 1.00
min_max_norm(data.frame(x = c(1, 2, 3), y = c(-10, 0, 10), id = c("a", "b", "c")))
#> x y
#> 1 0.0 0.0
#> 2 0.5 0.5
#> 3 1.0 1.0