使用sed删除所有不匹配模式的行

Question

Very new (meaning a day) to sed. I have worked a way to remove the lines that I do not need for an updated file with find/replace of a singular command

I have a list of objects as follows from a fortigate configuration:

config firewall address <--keep line
edit "item1"
next
edit "item2"
next
edit "item3"
next
edit "item4" <--keep line
unset associated-interface <--keep line and use as anchor for one line above and below
next <--keep line
edit "item5" <--keep line
unset associated-interface <--keep line and use as anchor for one line above and below
next <--keep line
edit "item6"
next
end <--keep line

I am trying to have it so when it is all said and done I am trying to keep the three lines (item4 & item5) and remove all of the other lines. Also, if possible keep the first and last line.

Answer 1

Your description lacks a great deal of precision, but here goes something :

sed -n -e '/^config/p;/item[45]/p;/^end/p' forti
config firewall address
edit "item4"
edit "item5"
end

Edit : this answer was given 2 hours before the edit that required the lines up to and following next after the item4 and item5 to be kept...

Answer 2

This might work for you (GNU sed):

sed '/config/,/end/!d;/config/b;/end/b;/item4/,+2b;/item5/,+2b;d' file

Delete all lines not between config and end . Print lines that begin config or end and lines that start with item4 or item5 and the following two lines.

Answer 3

This is probably not the best way to do it, but I couldn't find any other way around it.

I narrowed it down to four commands, which you can put in a bash script (replace file.txt with your own source file, and /tmp/finalFile with whatever destination you need):

head -n1 file.txt >> /tmp/finalFile
head -n $(echo $(grep -ni item6 file.txt | cut -d: -f1) - 1 | bc) file.txt > /tmp/tmpfile
tail -n $(echo $(wc -l /tmp/tmpfile | awk '{print $1}') - $(grep -n item4 /tmp/tmpfile | cut -d: -f1) + 1 | bc) /tmp/tmpfile >> /tmp/finalFile
tail -n1 file.txt >> /tmp/finalFile

When I ran the above commands, this is the final file in /tmp/finalFile :

config firewall address 
edit "item4" 
unset associated-interface  
next 
edit "item5" 
unset associated-interface 
next 
end 

Some explanation:

• The first line is simply the top line grabbed by head , appended to finalFile
• The second step is basically look for every line up to item5 , and save that in a temp file.
• The third step grabs that temp file, and calculates how many lines it needs to tail , to get all the lines up to item6 , not including it, and grabs those lines, appending the output to finalFile .
• Finally, the last line grabs the last line from the original file with tail , and appends that to finalFile