Hi All,
I was hopeing someone might be able to help me with a script I think using Sed that would find the file (ex.Nazs1) go inside this file and delete the lines that end in 88888888. And then make a new file with what is left after the lines have been deleted. Below is what is inside one of the files. I have about 50 files that need to be done. I would like to be able to just insert the filename into the script and then run it.

Mar 3 17:14 bin_prtrimester-20060303.1714.56.88888888
Mar 3 17:14 bin_prtrimester-20060303.1714.24.88888888
Dec 5 10:53 bin_safepr-20051205.1053.24.5790877
Nov 3 14:14 bin_safepr-20051103.1414.44.88888888
Nov 3 2004 bin_danna-20041103.1638.36.4057569
Nov 3 2004 bin_danna-20041103.1636.23.4057847
Nov 3 2004 bin_danna-20041103.1636.19.4057500
Nov 3 2004 bin_danna-20041103.1635.43.4057598
Nov 3 2004 bin_danna-20041103.1635.42.4057580
Nov 3 2004 bin_danna-20041103.1635.41.4057554
Nov 3 2004 bin_danna-20041103.1635.40.4057511
Nov 3 2004 bin_danna-20041103.1635.36.4057449
Nov 3 2004 bin_danna-20041103.1022.57.4057592
Oct 7 2004 bin_fxfever-20041007.1655.19.84900353
Sep 22 2004 bin_fxfountain-20040922.1808.12.88888888
Aug 3 2004 bin_fxfever-20040803.1746.53.88888888
Aug 3 2004 bin_fxfever-20040803.1745.37.88888888
Aug 3 2004 bin_fxfever-20040803.1444.16.88888888
Sep 10 2002 bin_show-20010613.1630.12.88888888


All my files are in the /tmp directory.

Thanks,

Nazs

Yes, sed would be the answer. Here is a sample of what sed does with the output you wanted. I shrank the output that you gave to save space:

So, the way I used sed is that it searches for all lines in file 'sample' and strips away anything that ends in '88888888'.

I'll leave it up to you to create the script . If you need help, paste the script that you have written so far and we'll help fix whatever issues you have.

-twantrd

You can use the -i option to edit the files in place.

sed -i '/\.88888888$/d' filename

You didn't provide information on how the file can be selected, such as the filename pattern. or if you need to search for files with the pattern. That would determine what would work best, a for log, or piping the output of a grep search.

find /tmp -iname "<pattern>" | sed -i '/\.88888888$/d'

or

for file in /tmp/<pattern>; do
sed -i '/\.88888888$/d' "$file"
done

Using a "grep" option, you can search for the "\.88888888$" pattern and return the names of the files containing the pattern.

easier one with grep itself,

what is the need for two process cat and sed and kernel ds,
just sed would do,

i missed out one more easier way in awk;