Skip to contents

customization

Source: vignettes/customization.qmd

Customization

This section covers customization options for tinytable. For format-specific customization, see:

library(tinytable)
options(tinytable_tt_digits = 3)
x <- mtcars[1:4, 1:5]

Methods to support non-data frame objects

The tt() function is a “generic” that can dispatch an object to an appropriate method based on the class of that object. The tinytable supplies methods for data.frame, data.table, and tibble. But users can define new method to automatically create tables for whatever object they like.

For example, let’s say we fit a linear regression model (class lm) and extract the results using the tidy() function from the broom package. We can create a table from the results as follows:

library(broom)
library(tinytable)
mod <- lm(mpg ~ hp + factor(cyl), data = mtcars)
results <- tidy(mod)
tt(results)
term estimate std.error statistic p.value
(Intercept) 28.65 1.5878 18.04 0.0000000000000000592
hp -0.024 0.0154 -1.56 0.1299540446968083796
factor(cyl)6 -5.968 1.6393 -3.64 0.001092088864954983
factor(cyl)8 -8.521 2.3261 -3.66 0.0010286170048313355

Alternatively, we can create an S3 method for the lm class, and then pass any model we fit to that class to automatically obtain a nice table.

tt.lm <- function(x, ...) {
  results <- broom::tidy(x)
  out <- tt(results, ...)
  out <- format_tt(out, j = "p.value", escape = TRUE,
    fn = \(p) format.pval(p, digits = 1))
  return(out)
}

lm(mpg ~ hp, data = mtcars) |> tt(digits = 1)
term estimate std.error statistic p.value
(Intercept) 30.1 1.63 18 <2e-16
hp -0.07 0.01 -7 2e-07 
lm(mpg ~ hp + factor(cyl), data = mtcars) |> tt(digits = 1)
term estimate std.error statistic p.value
(Intercept) 28.65 1.59 18 <2e-16
hp -0.02 0.02 -2 0.130
factor(cyl)6 -5.97 1.64 -4 0.001
factor(cyl)8 -8.52 2.33 -4 0.001

Shiny

tinytable is a great complement to Shiny for displaying HTML tables in a web app. The styling in a tinytable is applied by JavaScript functions and CSS. Thus, to ensure that this styling is preserved in a Shiny app, one strategy is to bake the entire page, save it in a temporary file, and load it using the includeHTML function from the shiny package. This approach is illustrated in this minimal example:

library("shiny")
library("tinytable")

fn <- paste(tempfile(), ".html")
tab <- tt(mtcars[1:5, 1:4]) |>
  style_tt(i = 0:5, color = "orange", background = "black") |>
  save_tt(fn)

shinyApp(
  ui = fluidPage(
    fluidRow(column(
      12, h1("This is test of tinytable"),
      shiny::includeHTML(fn)
    ))
  ),
  server = function(input, output) {
  }
)