delete specific character “[” from a file

Question

I want to remove a character "[" from a file. I tried

 sed -i 's/[//g' 'filename'

however I get the following error

sed: -e expression #1, char 6: unterminated `s' command

Answer 1

Yes, the [ character is special as it starts a [...] group (a bracketed expression).

With sed on OpenBSD, your command gives a more helpful error message:

$ sed 's/[//g'
sed: 1: "s/[//g": unbalanced brackets ([])

To delete all [ characters using sed, escape it:

sed -i 's/\[//g' file

Or put it inside a bracketed expression:

sed -i 's/[[]//g' file

Or, use tr,

tr -d '[' <file >file.new

Also, don't use in-place editing with sed until you know the expression that you are trying out actually works, or you will possibly have to restore your data from backups.

Answer 2

Since the question was originally tagged with awk and for no real good reason:

awk '{gsub(/\[/, "")}1' file >newfile

This will print the file with all [ removed and save it to newfile.

Using gawk you can make edits in place:

gawk -i inplace '{gsub(/\[/, "")}1' file

I wouldn't use either of these though unless you are also using awk to take other actions on the file.