how to use sed delete Unicode in some range?

问题:

I want to remove Unicode in some range, e.g.:

echo "abcＡＢＣ123" | sed 's/[\uff21-\uff3b]//g'

expect "abc123", but get:

sed: -e expression #1, char 20: Invalid range end

or use:

echo "abcＡＢＣ123" | sed 's/[Ａ-Ｚ]//g'

get:

sed: -e expression #1, char 14: Invalid collation character

回答1:

Unicode support in sed is not well defined. You may be better off using command line perl:

echo "abcＡＢＣ123" | perl -CS -pe 's/[\x{FF21}-\x{FF3B}]+//g'

abc123

It is important to use -CS flags here to be able to get correct UTF8 encodings for input/output/error.

回答2:

Not sure why sed is not working, but you can use tr instead

$ echo 'abcＡＢＣ123' | tr -d 'Ａ-Ｚ'
abc123

From man tr

tr - translate or delete characters

-d, --delete delete characters in SET1, do not translate