model_offset: Extract a model offset

In DavisVaughan/hardhat: Construct Modeling Packages

Description

model_offset() extracts a numeric offset from a model frame. It is inspired by stats::model.offset(), but has nicer error messages and is slightly stricter.

Usage

model_offset(terms, data)

Arguments

terms	A "terms" object corresponding to data, returned from a call to model_frame().
data	A data frame returned from a call to model_frame().

Details

If a column that has been tagged as an offset is not numeric, a nice error message is thrown telling you exactly which column was problematic.

stats::model.offset() also allows for a column named "(offset)" to be considered an offset along with any others that have been tagged by stats::offset(). However, stats::model.matrix() does not recognize these columns as offsets (so it doesn't remove them as it should). Because of this inconsistency, columns named "(offset)" are not treated specially by model_offset().

Value

A numeric vector representing the offset.

Examples

x <- model.frame(Species ~ offset(Sepal.Width), iris)

model_offset(terms(x), x)

xx <- model.frame(Species ~ offset(Sepal.Width) + offset(Sepal.Length), iris)

model_offset(terms(xx), xx)

# Problematic columns are caught with intuitive errors
tryCatch(
  expr = {
    x <- model.frame(~ offset(Species), iris)
    model_offset(terms(x), x)
  },
  error = function(e) {
    print(e$message)
  }
)

DavisVaughan/hardhat documentation built on Oct. 5, 2021, 9:53 a.m.