Introducing {numberwang} - numbers to words and vice versa)



numberwang will convert floating point numbers (and integers) to their word representations, and vice versa.

The key differentiator of this package, compared to {nombre}, is that it supports decimal representations by listing individual decimal digits.

See below for a comparison with {nombre}, {english}, {spanish} and {xfun}.

What’s in the box

  • num_to_words(num, decimals = 3) converts numeric vectors to character vectors
  • words_to_num(words) converts character vectors to numeric vectors


You can install from GitHub with:

# install.package('remotes')

That’s numberwang


c(12, pi, exp(10))
#> [1]    12.000000     3.141593 22026.465795
num_to_words(c(12, pi, exp(10)))
#> [1] "twelve"                                               
#> [2] "three point one four two"                             
#> [3] "twenty-two thousand and twenty-six point four six six"
  'negative fifteen point seven',
  "twenty-three point seven zero one",
  "ten million, two thousand and forty two"
#> [1]      -15.700       23.701 10002042.000


  • Floating point representation isn’t exact, which means that for very large numbers or very small numbers, and for lots of numbers inbetween, all the digits after the first few must often be taken with a grain of salt.
  • E.g. formatC(0.3-0.2, format = 'f', digits = 20) = 0.09999999999999997780
  • E.g. 1e25 should just be a 1 followed by zeros, but in floating point representation we get: formatC(1e25, format = 'f', digits = 0) = 10000000000000000905969664
  • The {english} and {nombre} packages will warn you about precision loss like this, but {numberwang} won’t.

Feature comparison

numberwang nombre english spanish xfun
function num_to_words nom_card english to_words n2w
working range double() double() +/- 2^90 [0,999] +/- 1e15
decimals Yes as fractions No No No
spanish No No No Yes No
words-to-numbers words_to_num nom_uncard (ints only) No No No

Just for fun - Let’s rotate the board!

#> [1] 1.942426e+130
  exp(300), format = 'f', digits = 0
#> [1] "19424263952412558251527551342607068708348876681949259689720520473778975990853476787505520791207218561168735878224980403355291484160"
  nineteen duoquadragintillion, four hundred and twenty-four
  unquadragintillion, two hundred and sixty-three quadragintillion, nine
  hundred and fifty-two novemtrigintillion, four hundred and twelve
  octatrigintillion, five hundred and fifty-eight septentrigintillion,
  two hundred and fifty-one sextrigintillion, five hundred and
  twenty-seven quintrigintillion, five hundred and fifty-one
  quattuortrigintillion, three hundred and forty-two tretrigintillion,
  six hundred and seven duotrigintillion, sixty-eight untrigintillion,
  seven hundred and eight trigintillion, three hundred and forty-eight
  novemvigintillion, eight hundred and seventy-six octavigintillion, six
  hundred and eighty-one septenvigintillion, nine hundred and forty-nine
  sexvigintillion, two hundred and fifty-nine quinvigintillion, six
  hundred and eighty-nine quattuorvigintillion, seven hundred and twenty
  trevigintillion, five hundred and twenty duovigintillion, four hundred
  and seventy-three unvigintillion, seven hundred and seventy-eight
  vigintillion, nine hundred and seventy-five novemdecillion, nine
  hundred and ninety octadecillion, eight hundred and fifty-three
  septendecillion, four hundred and seventy-six sexdecillion, seven
  hundred and eighty-seven quindecillion, five hundred and five
  quattuordecillion, five hundred and twenty tredecillion, seven hundred
  and ninety-one duodecillion, two hundred and seven undecillion, two
  hundred and eighteen decillion, five hundred and sixty-one nonillion,
  one hundred and sixty-eight octillion, seven hundred and thirty-five
  septillion, eight hundred and seventy-eight sextillion, two hundred and
  twenty-four quintillion, nine hundred and eighty quadrillion, four
  hundred and three trillion, three hundred and fifty-five billion, two
  hundred and ninety-one million, four hundred and eighty-four thousand,
  one hundred and sixty


  • R Core for developing and maintaining the language.
  • CRAN maintainers, for patiently shepherding packages onto CRAN and maintaining the repository