how to extract service names out of file and save them in clean version in file? [closed]

Question

i try to figure out how to extract all service names out of /etc/services and send them to text file sort them alphabetically remove any duplicates get rid of any blank lines and unnecessary spaces and have line count

Answer 1

The file /etc/services looks like this:

# Network services, Internet style
#
# (more comments)

tcpmux    1/tcp     # TCP port service multiplexer
echo      7/tcp
...
discard   9/tcp     sink null
discard   9/udp     sink null
...
ssh       22/tcp    # SSH Remote Login Protocol
...

That is, there are

• lines that consist only of a comment
• lines with trailing comments
• empty lines
• duplicates in the first column

To remove all of them, the following command is one possible way:

awk '!/^\s*$/ && !/^\s*#/ {print $1}' /etc/services | sort -u | cat -n

Output

 1  acr-nema
 2  afbackup
 3  afmbackup
 4  afpovertcp
 5  afs3-bos
 6  afs3-callback
 7  afs3-errors
...

Explanation

awk '!/^\s*$/ && !/^\s*#/ {print $1}' /etc/services

In general, awk reads a given file (or stdin) and then executes the following for each line:

condition_for_a_line { commands_when_condition_is_met }    

We use !/^\s*$/ && !/^\s*#/ as the condition which means line is not empty and line doesn't start with a comment. When both conditions are met, the first column $1 is printed. This gives us all first columns of the file without empty and comment lines.

sort -u

This sorts the data and eliminates duplicates in one go.

cat -n

Simply adds line numbers.

Answer 2

Here's my sed solution:

</etc/services sed '/^\w/!d;s/\s.*//' | sort -u >out

^{sed '…' file instead of <file sed '…' is possible but not preferable, as Stéphane Chazelas explains.}

This will first eliminate every line beginning with something else than a word character, then delete everything from every line beginning with the first whitespace character, then sort alphabetically and eliminate duplicate lines and finally redirect the output into the file out.

To get a line count of this file, I'd use wc:

<out wc -l

Explanations

• </etc/services – let bash open the specified file and assigns its content to stdin, it doesn't matter if this redirection is placed before or after the command, you could even do >out sed '…' <in if you like
• sed '…' – let sed execute the expressions …
• /^\w/!d – don't (!) delete every line beginning (^) with a word character, but every line that doesn't match this criterion
• s/\s.*// – in every line, substitute the first match of a whitespace character and everything (.*) that follows with nothing (= delete it)
• command1 | command2 – let bash assign command1's stdout to command2's stdin
• sort -u – let sort sort the lines alphabetically and output only the first of an equal run (-u or --unique)
• >out – let bash redirect the stdout to the specified file
• wc -l – let wc print the newline counts

Answer 3

awk '{print $1}' /etc/services | grep -v # | uniq | sort > yourFile