Shell Programming and Scripting

Or with awk

Code:

awk 'NR>3' myapplication.log

@Little
This:
sed '1,3d' myapplication.log > myapp;mv myapp myapplication.log
can be replaced by this:
sed -i '1,3d' myapplication.log
-i deletes it directly in the file, no need print, move etc

Dear Little and Jotne,

I don't think that this was the requirement, but thanks for the code. I will use those in other things where I've previously messed about with head & tail


Many thanks for smartening my code! It shows we can all learn.



Robin

Not sure about that:

I have around 800 lines which start with the followng three strings
PWRBRKER-3493
PWRBRKER-7834
SCHEDULER-ERROR

Yes, I suppose it depends on interpretation. Mine was that there were 800 or so unwanted lines in the file and each of them started with the given strings, and needed to be removed to leave the rest. I can see how it can also be read that there is an 800 line file and the first three records need to be removed.

Oh well, I suppose that between us, we've answered both.

omega3 can you confirm that you have an answer one-way or another?



Thanks,
Robin

thanks i didnot knew this option.

You should be careful when moving new files over old ones if the files permissions, metadata, inode number, etc. are important.

Code:

blah .... > newfile && cp newfile oldfile && rm newfile

sed -i does not do anything directly in the file. Just like using > and mv, it writes a new file which it then moves.

The most notable side-effect of using -i is that an otherwise portable script such as 1,3d is rendered unusable with most sed implementations. In such cases, -i is worse than worthless.

Regards,
Alister