Multibyte non-UTF-8 locales

locale.R

# ==== Creating UTF-8 strings ====
# This is how to create a string with UTF-8 encoding. This should work 
# regardless of the current locale settings.
x <- rawToChar(as.raw(c(0xe5, 0x8d, 0x88)))
Encoding(x) <- "UTF-8"
x
# [1] "午"

# Another string, 'Δ★😎'
pat <- rawToChar(as.raw(c(0xce, 0x94, 0xe2, 0x98, 0x85, 0xf0, 0x9f, 0x98, 0x8e)))
Encoding(pat) <- "UTF-8"
cat(pat)
# Δ★😎


# =======================
# Setting locale
# =======================

# By default, Mac and Linux use UTF-8 encodings, but sometimes it's useful to
# use a multibyte, non-UTF-8 locale for testing.

# ==== Mac ==== 
# On a Mac, you can use a UTF-8 locale like so. (This is default setting for US
# English).
Sys.setlocale("LC_ALL", "en_US.UTF-8")

# To use a multibyte non-UTF-8 locale:
Sys.setlocale("LC_ALL", "ja_JP.SJIS")


# ==== Ubuntu ====

# On Ubuntu, you need to enable a multibyte non-UTF-8 locale, like ja_JP.EUC-JP.

# As root, put the following in a new file /var/lib/locales/supported.d/ja
ja_JP.UTF-8 UTF-8
ja_JP.EUC-JP EUC-JP

# After adding that file, run:
sudo dpkg-reconfigure locales

# Restart R if it's already running. Then you can run the same R code as above,
# but with the ja_JP.EUC-JP locale:
Sys.setlocale("LC_ALL", "ja_JP.EUC-JP")