cat /etc/squid/squid.conf| sed -e 's/^#.*//g' | sed -e ':r;N;s/\n$//g;br'

First part:

sed -e 's/^#.*//g'

strips any line beginning (^) with # and anything that follows after the hash.

Next part:

sed -e ':r;N;s/\n$//g;br'

1. create register r,
2. put newline (N) in register r,
3. substitute all ending newlines (\n$) with nothing //,
4. repeat with register r.

Using bash:

cat test | sed -s 's/\([a-zA-Z0-9]\)/\1 /g'

Note that sed needs the grouping operators ( and ) escaped.

Suppose we have a text file containing the following lines:

number 1
number 3

and that we want to insert number 2 as the second line between number 1 and number 3, we would issue:

sed -i '2inumber 2' file.txt

where the parts of the fragments represent:

• 2 the line number,
• i for the insert command, and
• number 2 the data that has to be inserted.

which will modify the file in-place such that the contents of the file after the insertion becomes:

number 1
number 2
number 3

To print out all the lines between 5 and 10, issue:

sed -n '5,10p' file.txt

One way to normalize filenames is to pass the name through the streamline editor, sed:

echo "/path/to/file.txt" | LANG=C sed "s/[^\x20\x30-\x39\x40-\x5a\x61-\x7a\x2f\x5c\x5b\x5d]//g"

in order to remove unwanted characters.

The hexadecimal set that sed matches in the example corresponds to all characters 0 to 9, a to z, A to Z plus space, the forward- and backward slashes as well as the square brackets that are used typically to store the YouTube ID within the filename.