Replacing [0-9] with [A-J] not working with sed

Question

I am working on a command that replaces all digits 0-9 with their corresponding letters in sed. I know I'm doing something wrong, but sed is not interpreting the replacement regex as anything but a string literal.

The command I am using is sed -r 's/[0-9]/[A-J]/g' log > ~/output.txt

It seems pretty straightforward to me, but I have been stuck on it for about an hour. The output I receive just replaces 0-9 with the string "[A-J]"

Answer 1

In substitution, only the match (left-hand side) is a regular expression. The replacement is more or less just a literal string (with some backslash expansion) – it's not a regex.

You want transliteration, not substitution, so replace s with y:

echo 34031445 | sed 'y/0123456789/ABCDEFGHIJ/'

sed can't use ranges in y, but Perl can:

echo 34031445 | perl -pe 'y/0-9/A-J/'

Or just use tr:

echo 34031445 | tr 0-9 A-J