How to force uniq to distinquish between em- and en-dashes?

Question

uniq (GNU coreutils 8.5) does not seem to distinguish between em- and en-dashes:

$ echo -e "a–b\na—b" | uniq -c

  2 a–b

Is there any way to force this distinction? I've tried various settings for LC_COLLATE with no luck.

score 3 · Accepted Answer · answered Oct 28 '11 at 11:56

3

Worked for me

echo -e "a–b\na—b" | LC_COLLATE=C uniq -c
      1 a–b
      1 a—b

answered Oct 28 '11 at 11:56

I see. I was trying various permutations of `LC_COLLATE=en_GB.utf8` assuming that it _must_ be `utf8` to work. `LC_COLLATE=C` produces the expected results. Cheers! – Oct 28 '11 at 12:01

1 Answers1