HP-UX

Hi,

I am using HP-UX B 11.23

I want to copy files from one directory into another directory in such a way that, only the nonexistent files in target directory are copied from source directory

e.g.
dir1 has
file1
file2
file3
dir2 has
file1
file3
now I want that a command (non-interactive) to copy only file2 from dir1 to dir2

Thanks and Regards,
Chetana

Code:

cd dir1
 for i in `ls`; do  if [ -e $i ]; then echo "$i exists"; else echo "$i doesnot exist"; cp $i ../dir2/;  fi; done

HTH,
PL

Hi Daptal

Many Thanks for the reply

The option you suggested works.
However I am not sure if it will ipmact the performance, since the (exists) test will be carried for each file individually

This is like -exec
Do we have any option like xargs which will execute only once for this?

Initially I thought of 'diff' for comparing the files in directories but it compare contents of the files as well which is not feasible in my case since I have binary files of quite big size

These are Oracle archivelog files, each 150-200 Mb and large number of quantity

Also I tried following which too is slow

ls -ltr dir1>dir1FileList.txt
ls -ltr dir2>dir2FileList.txt
diff dir1FileList.txt dir2FileList.txt|awk '{print $9}'>reqdFiles.txt

Thanks and Regards,
Chetana

have you tried using diff on the directories directly, instead of sending listings to text files and processing that?

Code:

diff dir1 dir2

gives output like the following:

Code:

Only in dir1: file1
Only in dir1: file2
Only in dir2: file3

So you could do the following (assuming dir1 is the source and dir2 is the target):

Code:

for FILE in $(diff dir1 dir2|grep dir1 | cut -f4 -d' ')
do
  cp -p dir1/${FILE} dir2/${FILE}
done

I dont think it would be impacting the performance much .I am not quite confident abt this as i have not done benchmarking on this.

Also if the performance is really bad you can try this

ls -l dir1 | awk ' {print $5,$8}' | sort -k 2,2 > file1
ls -l dir2 | awk ' {print $5,$8}' | sort -k 2,2 > file2

If the file 1 in dir 1 is same as file 1 in dir2 dont bother checking the sizes just take the file name only and write a simple one liner to get the file names that donotexist

HTH,
PL

rsync can help you ?!

i am not sure, but i will suggest you to check the feasibility.

Hi.

You could also do something like:

Code:

false | cp -i $DIR1/* $DIR2

or

yes no | cp -i $DIR1/* $DIR2