Builds the data-bound dataset shared by the test factories. Column kinds are inferred from the column classes (integer/logical/factor/character -> discrete, double -> continuous); discrete factor/character columns are coded to integer codes before reaching the Rust core.

dataset(df)

Arguments

df

A data.frame (or object coercible to one).

Value

An object of class fastcit_dataset wrapping the bound dataset and its column names.

Details

Missing values are rejected: any NA (in a numeric column, a factor, or after coding a character column) reaches the core as NaN and errors with "missing data: ...". Choose a missing-data convention (e.g. na.omit(df)) before binding.

Examples

df <- data.frame(A = sample(0:1, 50, TRUE), X = rnorm(50))
data <- dataset(df)