In jamesdunham/concatenate: Human-Friendly Text from Unknown Strings

Functions in concatenate construct human-friendly text from elements whose values aren't known in advance, as in calls to message, warning or stop.

Basics

# rmarkdown::render("README.Rmd")
knitr::opts_chunk$set(
  echo = TRUE,
  collapse = TRUE,
  comment = "#>",
  fig.path = "README-",
  cache = FALSE)
library(concatenate)

Each function in concatenate returns a comma-separated string. (A length-one character vector.) The workhorse function in concatenate is cc.

cc("one fish", "two fish")

Its wrappers cc_or and cc_and insert "or" and "and" between the last two elements of the input. This is cleaner than wrestling with paste.

cc_and("this", "that", "the other")
cc_or("one way", "another")

cn and its derivatives combine these functions with sprintf-like substitution and the grammatical number awareness of ngettext.

Like ngettext, the cn functions return different strings depending on the number of the object, and two substitutions are made sprintf-style:

• "%c": the comma-concatenated values of the object, as in cc
• "%n": the number of the object

This encourages verbose user-facing messages from clean code.

x <- unique(iris$Species)
cn_and(x, "a single species: %c", "%n unique species: %c")

There are row-wise data.frame methods for the cn functions.

singular <- "%n row: %c"
plural <- "%n rows whose values are %c"
cn(chickwts[1, ], singular, plural)
cn_and(chickwts[1:3, 1, drop = FALSE], singular, plural)

The cc functions are also available as binary infix operators.

x <- "important value"
x %c% "!"

Install

concatenate is available via CRAN.

install.packages("concatenate")

Or, get it from GitHub.

devtools::install_github("jamesdunham/concatenate")

jamesdunham/concatenate documentation built on May 18, 2019, 11:19 a.m.