将目录名称附加到文件末尾并将其移动

Question

I am finding some files in a directory using this command:

find /Users/myname -type f

output is:

/Users/myname/test01/logs1/err.log
/Users/myname/test01/logs1/std
/Users/myname/test01/logs2/std
/Users/myname/test02/logs2/velocity.log
/Users/myname/test03/logs3/err.log
/Users/myname/test03/logs3/invalid-arg

I need to move this files to a different directory by appending the test directory name to the end of the files. Like below:

err.log-test01
std-test01
std-test01
velocity.log-test02
err.log-test03
invalid-arg-test03

I am trying with the cut command but not getting the desired output.

find /Users/myname -type f | cut -d'/' -f6,4

plus, I also need to move the files to a different directory. I guess a suitable way could be there using sed command, but I am not proficient with sed. How this can be achieved in an efficient way?

Answer 1

You can let find create the mv command, use sed to modify it and then have it run by the shell:

find /Users/myname -type f -printf "mv %p /other/dir/%f\n" |
           sed 's,/\(test[0-9]*\)/\(.*\),/\1/\2-\1,' | sh

This assumes there are no spaces in any argument, otherwise liberally add ' or " . Also run it without the final | sh to see what it actually wants to do. If you need to anchor the test[0-9]* pattern better you can include part of the left or right string to match:

's,myname/\\(test[0-9]*\\)/\\(.*\\),myname/\\1/\\2-\\1,'

Answer 2

You can move it from the dst to the dst_dir appending the directory, using awk , and the target name would be awk -F/ '{print $5 "-" $4}' . The full command could be as simple as:

for i in `find . -type f`
   do mv $i /dst_dir/`echo $i| awk -F/ '{print $5 "-" $4}' `
done

Answer 3

There are a number of things going on that you may want to use a helper script with find to insure you can validate the existence of the directory to move the files to, etc.. A script might take the form of:

#!/bin/bash

[ -z $1 -o -z $2 ] && { # validate at least 2 arguments
    printf "error: insufficient input\n"
    exit 1
}

ffn="$1"                # full file name provided by find
newdir="$2"             # the target directory

# validate existence of 'newdir' or create/exit on failure
[ -d "$newdir" ] || mkdir -p "$newdir"
[ -d "$newdir" ] || { printf "error: uname to create '$newdir'\n"; exit 1; }

tmp="${ffn##*test}"     # get the test## number
num="${tmp%%/*}"

fn="${ffn##*/}"         # remove existing path from ffn

mv "$ffn" "${newdir}/${fn}-test${num}"  # move to new location

exit 0

Save it in a location where it is accessible under a name like myscript and make it executable (eg chmod 0755 myscript ) You may also choose to put it in a directory within your path. You can then call the script for every file returned by find with:

find /Users/myname -type f -exec ./path/to/myscript '{}' somedir \;

Where somedir is the target directory for the renamed file. Helper scripts generally provide the ability to do required validation that would otherwise not be done in one-liners.