help with grep

Ave3 · 11-08-2014, 02:38 PM

Hi guys,

I have one sort of file with the same name in many sub-directories in the same mother folder, I need to pull out the last line in each file of those files I don't know how to do that in once, it takes very long time to go to each sub directory and in to the file and (tail -1) for each one to take the last line, I tried with grep command but I failed, any idea please?

Best,
Ave3

John VV · 11-08-2014, 03:49 PM

Quote:

I have one sort of file

What sort of file is this ?

and what have you tried so far ?

please post an example

syg00 · 11-08-2014, 03:52 PM

Perhaps the easiest way to print the last line of a file is to use

Code:

sed -n '$p' input.file

To find all the files, use the "find" command - it has an "-exec" option that will enable you to run the sed on each file found.
That should do it all for you - the hard part is learning the "find" parameters. See how you go.

Ave3 · 11-08-2014, 05:51 PM

Hi syg00,I am new using linux and never used find but after u told me I read about it and found the solution for my problm with this beautiful line ( find . -iname 'fe.dat' -exec tail -n1 {} \

, can you tell me what this bracket does or this part {} \; of the line?

Thank you johnvv for your concern.

Best,

syg00 · 11-08-2014, 06:25 PM

Good for you figuring it out.

Quote:

Originally Posted by manpage

The string `{}' is replaced by the current file name being processed everywhere it occurs in the arguments to the command

The semi-colon marks the end of the command being exec'd, and needs to be escaped (the backslash) so the shell doesn't try to interpret it.

Ave3 · 11-09-2014, 02:37 PM

Thank you that helped me.
I have another question, is there any way to run a script on many directories at once?

fatmac · 11-10-2014, 07:05 AM

That script should find all the files you have/want, run it from your home directory.
(find . -iname 'fe.dat' -exec tail -n1 {} \)

Ave3 · 11-10-2014, 01:13 PM

Yes this script is to find the files and it helped my first problem, but now my other concern is to run another script on many directory, I have many directories in one main directory what I do is open each directory and run a specific script that I have to get the output files, it takes a long time to open each directory and run the script in it.
My question was is it possible to find the directories and run the script on them at the same time.

makyo · 11-10-2014, 09:23 PM

Hi.

Commands xargs and parallel will keep a number of processes running. So if you find the paths of the directories, either one of those can then create a queue of processes filled until it runs out of work to do. The man page for xargs should get you going, and parallel may be in your repository (it's in Gnu/Debian jessie, for example).

Best wishes ... cheers, makyo

John VV · 11-10-2014, 10:22 PM

Quote:

Thank you johnvv for your concern.

it is not concern

it is getting the answers to the game of 20 questions answered

for starters
#1 ) is this "fe.dat'" file a text file or a binary file ?

#2 ) if a text file what format is it ?
CSV or what