a file with lines (aa ac a-b) unix sorts to (aa a-b ac) instead of the expected (a-b aa ac)
it is as if sort is ignoring the '-' character.
interestingly, a dash by itself is sorted correctly (a b c -) sorts to (- a b c).
why? anyway to change this behavior?
The sort order behaviour of sort(1) is controlled by your locale settings (see man locale).
There are a number of different locale settings, for example:
$ locale
LANG=en_US.UTF-8
LC_CTYPE="en_US.UTF-8"
LC_NUMERIC="en_US.UTF-8"
LC_TIME="en_US.UTF-8"
LC_COLLATE="en_US.UTF-8"
LC_MONETARY="en_US.UTF-8"
...
LC_ALL=
To choose the desired sort behaviour, you need to choose the correct LC_COLLATE value. In this case, the standard built in C (POSIX) locale is suitable:
$ sort testcase
aa
a-b
ac
$ LC_COLLATE=C sort testcase
a-b
aa
ac
If you prefer, you can set all the locale settings (thus being more consistent) by setting LC_ALL=C. Since these are environment variables, you can permanently set your sort order, with export LC_ALL=C or similar, in your shell startup script.
setting environment variable LC_ALL=C changes the behavior of sort. The default locale sort order must be treating '-' specially.