Shell Programming and Scripting

Hi All,
I am a newbie to Shell scripting. Please help me with the Following problem,

1. I have two files with the same name in different locations in the same machine.
Eg: /root/testfolder/a ---- location 1
/tmp/testfolder/a ----- location 2

2. I want to compare the files in location 1 and location 2.
Eg: diff /root/testfolder/a /tmp/testfolder/a
(Please suggest the best way using diff )

3. I must append only the lines missing in location 1(i.e. /root/testfolder/a) without losing any content of location 1.

Eg:
File in location 1/root/testfolder/a)

Added line in Location 1
This is a test Script
But i am stuck.Please help
This is Added line to LOCATION 1(/root/testfolder/a)

File in location 2/tmp/testfolder/a)
Added Line in the Location 2 (/tmp/testfolder/a)
This is a test Script
But i am stuck.Please help
This is added line for testing in LOCATION 2(/tmp/testfolder/a)


Executing diff command:

diff /root/testfolder/a /tmp/testfolder/a

Say output of the diff command looks like this:

1c1
< Added line in Location 1 (/root/testfolder/a)
---
> Added Line in the Location 2 (/tmp/testfolder/a)
4c4
< This is Added line to LOCATION 1(/root/testfolder/a)
---
> This is added line for testing in LOCATION 2(/tmp/testfolder/a)

Here,

I am concerned only to append those lines starting with ">" to the correct lines of /root/testfolder/a

Expected result in /root/testfolder/a

Added Line in the Location 2 (/tmp/testfolder/a)
Added line in Location 1
This is a test Script
This is added line for testing in LOCATION 2(/tmp/testfolder/a)
But i am stuck.Please help


Please help me develop a shell script to solve the above posted problem.
Thanks in advance
Karthick.

Is the order of the lines important? Or can you just add the missing lines to the end of the file? If so, you could simply

Code:

fgrep -vxf file2 file1 >>file2

assuming you have an fgrep which understands the -f option.

yes the order of files is important. i want to compare perl source code files.


Thanks,
Karthick

Have you tried diff3? Do you have access to a common descendant which has none of the changes from either file?

I would suggest diff -u and patch but it's probably much easier if you can work against an earlier version of the file.

I can't really understand patch command... and how it can help me. ....Plz explain me how i can use it?

I also dint understand about "working against earlier version of file"

Let's say we have a file. Call it "version 1". Then you copy it to another place, and make some edits. Call this "version 2a". Then copy it to yet another place, and make some different edits. Call this "version 2b".

Now, the simple way to merge changes from 2a to 2b is to find the difference between 1 and 2a, and apply that to 2b.

Code:

diff -u version1 version2a >diff2a
cd ../path/to/2b
patch -p0 <../back/to/diff2a

The patches produced by diff have absolute line numbers etc so you cannot blindly apply them. patch is able to introduce "fuzz" so it can squeeze a patch into place even if some changes have happened in the file.

If there are conflicts, of course, you still need to resolve them.

A version control system probably helps if you need to do this often.