Shell Programming and Scripting

Hi,

I have a file:
ABESCAT Bernard 88409 8813 8808 1 P3 DEL 3.0.4 S3.01M
Admin. national Base Pro1 99001 9901 9901 0 P4 ADM 3.2.0.1 41027
ADMINISTRATEUR PROD EST 09999 9999 9999 0 P4 ADM 3.0.4
ADRIANT Thierry 51004 5105 5104 1 P3 DEL 3.0.4 S3.01M
ALARCON Pierre Yves 34007 3408 3407 0 P3 DEL 3.0.4 S3.00g
ALARCON Pierre Yves 34011 3413 3413 1 P3 DEL 3.0.4 S3.01M
ALMENDROS Michel 44006 4457 4406 1 P3 DEL 3.0.4 S3.01M
AMSELLAM Rene 44011 4456 4411 1 P3 DEL 3.0.4 S3.01M
ANDRIEUX Evelyne 34004 3404 3404 1 P3 DEL 3.0.4 S3.01M
BEAUVILLAIN Jean-Mar 52002 5258 5258 0 P3 DEL non actif
BELLOIS Fabien 12408 1212 1202 1 P3 DEL 3.0.4 S3.01M
BELLON Thierry 44013 4451 4412 1 P3 DEL 3.0.4 S3.01M
BENET Arnaud 47009 4736 4706 1 P3 DEL 3.0.4 S3.01M


and I want only recupere XXXX X where X is a number.
I can not use awk because it is not always a the same position.
It is similar for cut. I think I should use sed but it is difficult for me...

Thanks

Tested on your sample data

Code:

sed 's/^.*[ ]\([0-9]\{4\}[ ][0-9]\)[ ].*$/\1/' sample

Gives

Code:

8808 1
9901 0
9999 0
5104 1
3407 0
3413 1
4406 1
4411 1
3404 1
5258 0
1202 1
4412 1
4706 1

Cheers
ZB

if the pattern of the numbers is always the same and there are no other occurences of this pattern u could try the code below, which i came across @ this web page http://pegasus.rutgers.edu/~elflord/unix/sed.html


grep "[:digit:]\{4\}[ -]\?[:digit:]\{1\}" file

hope this helps

Trouble is that'll print the entire line, which is not what the OP wants.

Using \(pattern\) within the search pattern and \1 for the replacement in a sed substitution expression allows us to capture only a certain portion of the matched pattern for output.

Cheers
ZB

Always you zazzybob! thanks!

I don't know how it work but your sed is Ok
It is very difficult to understand

Thinks for your link w33man. If you have other links, it should be excellent!