Copying files with specific extension from a list (text file) of directories

Question

I have a text file with list of certain directories that I want to copy *.xlsx files from them to another directory.

This is how the the text file (list.txt) is arranged:

PT_NAK01, PT_NAK04, PT_NAK05, PT_JAR03

What I have so far:

@echo off
set main_folder="\\internal.company.com\project folder\"
set my_folder="C:\_M__\files"
for /f "tokens=*" %%i in (list.txt) DO (
    xcopy "%main_folder%\%%i" "%my_folder%"
)

So the folders that I want to look into would be \\internal.company.com\project folder\PT_NAK01 etc.

What I don't know is how to pass the specific extension *.xlsx to this command.

Note: I haven't used /S switch with xcopy deliberately because I do not want the files in the sub-directories.

P.S. Solutions in powershell or cygwin work for me as well.

Answer 1

One thing that made an error in my example was how I set the text file with the folder names. It should be set up with carriage return as separator instead of comma-separated entries.

PT_NAK01
PT_NAK04
PT_NAK05

etc.

With that, this batch-file (in reference to MatSnow's and shellter's comments) works fine for the purpose of the question.

@echo off
set main_folder="\\internal.company.com\project folder\"
set my_folder="C:\_M__\files"
for /f "tokens=*" %%i in (list.txt) DO (
    xcopy "%main_folder%\%%i\*.xlsx" "%my_folder%"
)

Note: If you want to type this directly into the command line, you don't need double % for the variables.

Answer 2

This is a cygwinshell answer (bash is an advanced shell that should be reserved for when standard Posix shell (/bin/sh) is insufficient). Note that slashes are reversed intentionally.

I see the format in your list.txt is delimited with commas and whitespace. I am going to assume that this is literal and the reason none of what you've tried so far works. Therefore, I am parsing it with the explicit assumption that comma and then space (, ) is a delimiter and that there is no way to escape them (e.g. if you have a file named apples, oranges.txt then my code would erroneously parse files named apples and oranges.txt).

#!/bin/sh
main_folder="${1:-//internal.company.com/project folder}"
my_folder="${2:-c:/_Masoud/files}"
cd "$main_folder" || exit $?
IFS=', ' find $(cat list.txt) -maxdepth 1 -name \*.xlsx |while IFS= read xlsx; do
  mkdir -p "$my_folder/${xlsx%/*}"
  cp -a "$xlsx" "$my_folder/$xlsx"
done

I've done some extra work for you to make this more abstract. $main_folder is taken from your first argument (a missing argument will default to //internal.company.com/project folder) and $my_folder is taken from your second argument (if missing, it defaults to c:/_Masoud/files). Don't forget to quote your command-line arguments if they contain spaces or interpretable characters.

After determining your source and destination, I then try to change directories to the source directory. If this fails, the script will stop with the same exit code.

Now for the loop. I've changed the Input Field Separator ($IFS) to be the comma and space (, ) we talked about earlier and then glued the contents of list.txt into its arguments, followed by the requirement of being one level deep (to include PT_NAK05/foobar/baz.xlsx, use -maxdepth 2 or just remove that clause altogether to view the file tree recursively), followed by the requirement of a name matching *.xlsx (this is escaped because your shell would otherwise assume you're talking about the local directory). The output of this is then read into a loop line by line as $xlsx. We recreate the target's parent directory in the new target destination if it's not already present, then we copy the file to that location. cp -a preserves permissions and time stamps.