shrink: Subset only required columns

In hardhat: Construct Modeling Packages

Subset only required columns

Description

shrink() subsets data to only contain the required columns specified by the prototype, ptype.

Usage

shrink(data, ptype)

Arguments

data	A data frame containing the data to subset.
ptype	A data frame prototype containing the required columns.

Details

shrink() is called by forge() before scream() and before the actual processing is done.

Value

A tibble containing the required columns.

Examples

# ---------------------------------------------------------------------------
# Setup

train <- iris[1:100, ]
test <- iris[101:150, ]

# ---------------------------------------------------------------------------
# shrink()

# mold() is run at model fit time
# and a formula preprocessing blueprint is recorded
x <- mold(log(Sepal.Width) ~ Species, train)

# Inside the result of mold() are the prototype tibbles
# for the predictors and the outcomes
ptype_pred <- x$blueprint$ptypes$predictors
ptype_out <- x$blueprint$ptypes$outcomes

# Pass the test data, along with a prototype, to
# shrink() to extract the prototype columns
shrink(test, ptype_pred)

# To extract the outcomes, just use the
# outcome prototype
shrink(test, ptype_out)

# shrink() makes sure that the columns
# required by `ptype` actually exist in the data
# and errors nicely when they don't
test2 <- subset(test, select = -Species)
try(shrink(test2, ptype_pred))

hardhat documentation built on June 22, 2024, 10:06 a.m.