Build Formulas Using Tidy Selection Helpers.
tidyformula
tidyformula() translates formulas containing tidyselect-style selection helpers, expanding these helpers by evaluating dplyr::select() with the relevant selection helper on a supplied data frame.
Installation
You can install the development version of tidyformula from GitHub with:
# install.packages("devtools")
devtools::install_github("damian-t-p/tidyformula")
Example
We demonstrate how to build formulas from the variables of the following toy data frame:
library(tidyformula)
df <- data.frame(
x1 = rnorm(5),
x2 = rnorm(5),
x3 = rnorm(5),
y = rnorm(5)
)
The simplest usage is adding a selection of variables. The tidy-selected variables can be combined with other variables in the formula:
tidyformula(y ~ starts_with("x") + z, data = df)
#> y ~ x1 + x2 + x3 + z
#> <environment: 0x00000202ad6c3730>
The selection helper can have additional arguments, as with num_range
tidyformula(y ~ num_range("x", 1:2) + z, data = df)
#> y ~ x1 + x2 + z
#> <environment: 0x00000202ad6c3730>
When the selection helper appears as the first argument of a function, that function is distributed across the sum of the selected variables.
This works with single-argument functions
tidyformula(y ~ log(contains("x")), data = df)
#> y ~ log(x1) + log(x2) + log(x3)
#> <environment: 0x00000202ad6c3730>
as well as multiple-argument ones.
tidyformula(y ~ poly(contains("x"), 3), data = df)
#> y ~ poly(x1, 3) + poly(x2, 3) + poly(x3, 3)
#> <environment: 0x00000202ad6c3730>
The functions +, -, *, and ^ are not distributed by default.
tidyformula( ~ everything()*z + starts_with("x")^2, data = df)
#> ~(x1 + x2 + x3 + y) * z + (x1 + x2 + x3)^2
#> <environment: 0x00000202ad6c3730>
This behaviour can be overwritten with the nodistribute argument, which is a character vector of functions that should not be distributed.
tidyformula( ~ everything()*z + starts_with("x")^2,
data = df,
nodistribute = c("+", "-"))
#> ~x1 * z + x2 * z + x3 * z + y * z + (x1^2 + x2^2 + x3^2)
#> <environment: 0x00000202ad6c3730>