Insert a file text into another file after particular line

bloodstreetboy · 02-13-2013, 03:42 AM

I have a file 'install'.
I want to insert 8 to 18 lines of this file into another file 'space' after 17 lines.

I do not want to use shell script. I have done this already.

If you tell me using sed or head/tail in one line (So I can use it on terminal), it would be more appreciated.

1) If I want to insert complete file into another file at bottom, I use
$ cat 'file1' >> file2

2) If I want to insert particular text(suppose line number 7 to 20) of a file into another file at bottom, I use
$ tail -n +7 space | head -n 14 >> install

But what if I want to insert line number 7 to 20 of a file into another file at line number 8?
Please respond.

shivaa · 02-13-2013, 03:54 AM

Awk could do that:

Code:

~$ awk 'NR>=8 && NR<=18' install

OR, To append it's output at the end of 'space'
~$ awk 'NR>=8 && NR<=18' install >> space

But let's know how many lines space file already have?

bloodstreetboy · 02-13-2013, 04:12 AM

File 'space' has more than 500 lines.
I want to copy line number 7 to line number 20 of 'space' into file 'install' after line number 7.
It means line number 8 to 21 of install will be same as line number 7 to 20 of file 'space'. After that line number 22 of install will be same as old install's line number 8, line number 23 of install will be same as old install's line number 9 and so on.
If you are not getting this. I will tell you an example in next comment (if you want)

jpollard · 02-13-2013, 04:18 AM

Check the "ed" editor. Stupidly simple.

You can even use diff to compair two files and let it generate the necessary commands.

druuna · 02-13-2013, 04:23 AM

Have a look at this one-liner:

Code:

awk '{ print } NR==7 { while (( "sed -n '7,20p' space" | getline ) > 0 ) print }' install

The green part represents the line to insert, the blue part are the lines to be inserted.

An example run on two small files:

Code:

 $ awk '{ print } NR==8 { while (( "sed -n '3,6p' file_two" | getline ) > 0 ) print }' file_one 
1 - one
2 - one
3 - one
4 - one
5 - one
6 - one
7 - one
8 - one
3 - two
4 - two
5 - two
6 - two
9 - one

$ cat file_one
1 - one
2 - one
3 - one
4 - one
5 - one
6 - one
7 - one
8 - one
9 - one
$ cat file_two
1 - two
2 - two
3 - two
4 - two
5 - two
6 - two
7 - two
8 - two
9 - two

fortran · 02-13-2013, 04:36 AM

@druuna
Thanks, this looks correct output.
But can we not do this using sed command only?
OP wants to do this using sed command.

druuna · 02-13-2013, 04:48 AM

Quote:

Originally Posted by pavi_kanetkar

@druuna
Thanks, this looks correct output.
But can we not do this using sed command only?
OP wants to do this using sed command.

The OP did mention sed/head/tail, but experience learns that people ask about the commands they have tried (and failed) with. They tend to "ignore" the other/better solutions, mostly due to inexperience with those commands.

It might be possible to create a complicated sed script that can do this, but a script isn't wanted by the OP.

fortran · 02-13-2013, 04:50 AM

Great... Thanks Druuna !!!

bloodstreetboy · 02-13-2013, 05:06 AM

Sorry but I do not want to use `shell script` as well as `awk` ( sorry I did not mention this).
I have given you example of head/tail. I want something similar like this.
Can't we do this using head & tail and >> ?

druuna · 02-13-2013, 05:25 AM

Quote:

Originally Posted by bloodstreetboy

Sorry but I do not want to use `shell script` as well as `awk` ( sorry I did not mention this).

sed, head, tail, awk, etc are all normal commands (which can also be used in a full blown script). Why the limitation?

Quote:

I have given you example of head/tail. I want something similar like this.
Can't we do this using head & tail and >> ?

Probably. But it won't be pretty, will be "unreadable" and will use multiple temporary files. In other words: messy.see post #12

You mention that you already have a script that does what you want/need, why not use that from the terminal?

druuna · 02-13-2013, 05:32 AM

Using sed and cat:

Code:

sed -n '1,8p' file_one > /tmp/x.1 ; sed -n '9,$p' file_one > /tmp/x.2 ; sed -n '3,6p' file_two > /tmp/y.1 ; cat /tmp/{x.1,y.1,x.2} > file_new

file_new now contains:

Code:

$ cat file_new
1 - one
2 - one
3 - one
4 - one
5 - one
6 - one
7 - one
8 - one
3 - two
4 - two
5 - two
6 - two
9 - one

colucix · 02-13-2013, 06:01 AM

Quote:

Originally Posted by bloodstreetboy

But what if I want to insert line number 7 to 20 of a file into another file at line number 8?

Code:

sed -i 8r<(sed -n 7,20p infile) oufile

bloodstreetboy · 02-13-2013, 08:15 AM

Thanks colucix... I was asking something like this.

If I use 8r, it starts inserting from line number 9, I want it from line number 8 so I changed it to 7r.

Perfect.

Thank you..

colucix · 02-13-2013, 08:20 AM

Well done. You're welcome!