I have a small problem @ work. I am trying to write a script that will seach a directory for a file with the contents 'IS314.1'. Then move that file 2 a new directory. This is what I have but I am not sure I am using the right approach. (I am using SH because that is all I am familiar with.) ls -l | sort +4 | sort +5 > tempfile.txt cat tempfile.txt | while read file do filename=${file} echo $filename if $filename = uv* then what? I am not sure I am takign the right approach. Basically I need to find string IS314.1 in a file uv* and cp file to another directory as IS312.1 Please help!

I know I am doing something wrong but so far I have modded what I have above to: 1. wrote a script to do: ls -t uv* > listspl.txt 2. wrote a script to run after the main program runs (the one that generates IS314.1 - this cuts down on the files we need to look at) ls -t uv* > listspl2.txt diff -n listspl.txt listspl2.txt > listspl3.txt cat listspl3.txt | grep uv > listsplfin.txt 3. This part is breaking, I know something is feshmekeled but I can't figure out what I am doing wrong. cat /u1/uvspool/listsplfin.txt | while read file do filename=${file} cat /u1/uvspool/$filename | while read test do testme=${test} if $testme = "IS314.1" then cp /u1/uvspool/$filename /u2/dbms/MHCDATA/DWNLD/IS314.1 fi done done

oh yea and the "IS314.1" string will be found within the first 10 lines of the file so it is not necessary to continue beyond that.

Ok I think I have step3 working now but if someone has a better (i.e. more ellegant solution) please let me know - This is what I have: cat /u1/uvspool/listsplfin.txt | while read file do filename=${file} if(grep IS314.1 /u1/uvspool/$filename) then cp /u1/uvspool/$filename /u2/dbms/MHCDATA/DWNLD/IS314.1 break fi done cd /u1/dbms/LIVE.MHC exit 0

#!/bin/sh for file in `find . -exec grep -l IS314.1 {} \; ` ; do cat $file >> /u2/dbms/MHCDATA/DWNLD/IS312.1 # you may have more than one occurance. # should this overwrite or append? done

actually I would like to keep the most reacent version. (mtime comes ot mind - but where do you put the command?) - There will only be one version on the day the script is run. Also why do yo need the {}\;`; This is all new to me so please explain. Also why grep -l? I can't believe that you took all those lines and got it done in 2! I knew there was a better way but I am not that good at putting the commands togethers yet. Thanks, Mano

grep -l will print a file name where a match occurs and not the contents of the file. You can use this to generate a list of files that contain a substring you are looking for. A find command that does more than the default print needs to be terminated bye "\;". The "{}" represent the file matched by find. To find and keep the most recent file, the find command would need to be piped through sort and head -1. I'll work on the details an post more of an answer. If there is only one version when the file is run you can do a cp instead of cat. cp $file /u2/dbms/MHCDATA/DWNLD/IS312.1