How to replace a number of character with a single one?

Question

If the character is met more than once I must replace it with a single one. For example i have this in a file: ????aca, I must replace it with ?aca. I've tried it with tr, but didn't make it to work proper.

Answer 1

It seems easy with sed using backreferences:

sed 's/\(.\)\1\+/\1/g' infile

Answer 2

If you want to truncate consecutive runs of one particular character, tr -s does that.

tr -s '?' <file >file.new

Answer 3

Based on your limitation to ?

sed 's/[?]\{2,\}/?/g' YourFile

also POSIX compliant