How to write to first n lines of file in bash

Question

I have a file test.txt with following content

first
second
AAA
BBB
CCC
DDD

And I want to remove the first two lines and add new values to the first two lines,

so,

once the first two lines are removed, file should look like this:

AAA
BBB
CCC
DDD

And then should add the two values to first line and then the second line, so the file would finally look like below:

new value at line 1
new value at line 2
AAA
BBB
CCC
DDD

So I tried the below command, but how can I remove the first two lines?

SERVER_HOSTNAME=$(hostname)
SERVER_IP=$(ip -o route get to 8.8.8.8 | sed -n 's/.*src \([0-9.]\+\).*/\1/p')
sed -i "1s/.*/$SERVER_HOSTNAME/" /tmp/test.txt
sed -i "2s/.*/$SERVER_IP/" /tmp/test.txt

My problem is, when I first remove the first two lines, and execute the above command, it will replace the line number 1 and two with new values, but I want to add them on the top so the others (already existing content will shift down) will go down.

Answer 1

Could you please try following, written and tested with shown samples in GNU awk . This will edit 1st and 2nd lines with shell variable values and do an inplace save into Input_file itself.

In case you want to keep 1st line's content as well as print current content then one could remove next in FNR==2 OR FNR==1 conditions.

SERVER_HOSTNAME=$(hostname)
SERVER_IP=$(ip -o route get to 8.8.8.8 | sed -n 's/.*src \([0-9.]\+\).*/\1/p')
awk -v server_ip="$SERVER_IP" -v server_hostname="$SERVER_HOSTNAME" '
FNR==1{ print server_hostname; next}
FNR==2{ print server_ip;     ; next}
1
' Input_file > temp && mv temp Input_file

Explanation for above solution:

awk -v server_ip="$SERVER_IP" -v server_hostname="$SERVER_HOSTNAME" '
##Starting awk program from here, creating server_ip and server_hostname vars with respective shell vars.
FNR==1{ print server_hostname; next}
##Checking condition if this is 1st line then print server_hostname.
FNR==2{ print server_ip;     ; next}
##Checking condition if this is 2nd line then print server_ip here.
1
##1 will print current line.
' Input_file
##Mentioning Input_file name here.

Answer 2

Don't run sed -i repeatedly. Instead, combine all your commands into one script.

sed -i "1s/.*/$(hostname)/
    2s/.*/$(ip -o route get to 8.8.8.8)/
    2s/.*src \([0-9.]\+\).*/\1/" /tmp/test.txt

This is rather brittle, though; in particular, it will break if either of the command substitutions produces a slash in their output.

(IIRC ip has options to produce machine-readable output; you should probably look into that instead of replacing out the parts you don't want.)

If you want to add, then delete, the sed d and a commands do that, respectively. But removing two and adding two is obviously equivalent to replacing two.

sed -i "1d
1a\hello
2d
2a\hello" file

Unfortunately, the a command is poorly standardized, and it's unclear how two s commands would not work, so I'm leaving this as a sketch.

Answer 3

Your existing sed already does what you want; it doesn't need you to remove the first two lines yourself first.

$ cat tmp.txt
first
second
AAA
BBB
CCC
DDD
$ SERVER_HOSTNAME=example.local
$ SERVER_IP=127.0.0.1
$ sed -i "1s/.*/$SERVER_HOSTNAME/;2s/.*/$SERVER_IP/" tmp.txt
$ cat tmp.txt
example.local
127.0.0.1
AAA
BBB
CCC
DDD

Answer 4

You may use this gnu sed :

gsed -i -e "1i\\$SERVER_HOSTNAME\n$SERVER_IP" -e '1,2{d;q;}' /tmp/test.txt

Or using POSIX sed:

sed -i.bak "1,2{d;q;};3i\\
$SERVER_HOSTNAME
3i\\
$SERVER_IP
" /tmp/test.txt