Problem in using Sed to remove leading and trailing spaces

Question

Hi all,

I am using the following code to remove both leading and tailing spaces from all lines of a file A.txt

sed 's/^[ \t]*//;s/[ \t]*$//' ./A.txt > ./B.txt

The problem occurs on the lines where there is a t in the beginning or at the end. So say for example, the original line that starts with the string "timezone" becomes "imezone"

Can you please tell me what is happening here? and also if there is a known solution to the problem.

Thanks in advance.

Answer 1

You might have better luck if you just use the whitespace character class:

sed -e 's/^\s*//' -e 's/\s*$//' A.txt > B.txt

Answer 2

newer versions of sed understand \t but older versions may not have. You may be better off replacing \t with a literal tab (CTRL-V will give you that)

What version of sed are you using on what system? (sed --version for GNU sed)

Answer 3

Some older versions of sed don't understand C-style escape characters such as \t and treat that as two characters, '\' and 't'. You can avoid the problem by using a literal tab character -- if you're typing this directly into the shell, type Ctrl+V Tab.

Answer 4

Another alternative, using whitespace character classes, if your variety of sed doesn't support \s:

sed 's/^[[:space:]]*//;s/[[:space:]]*$//'

Answer 5

gawk

awk '{gsub(/^[[:space:]]+|[[:space:]]+$/,"")}1' file