在使用tar提取文件时，如何在通配符上使用OR运算符

Question

While I can find all the .tgz files within a folder and then extract only PDF, EPUB and MOBI files from it if it is present in the archive.

find '/home/pi/Downloads/complete/' -type f -name "*.tgz"| while read i ; do tar -xvzf "$i" -C /home/pi/Downloads/complete/ebook/ --strip=1 --wildcards --no-anchored '*.pdf' '*.mobi' '*.epub'; done

This line of code works perfectly when either of pdf, mobi or epub is present in the archive. However with this code, whenever there is no pdf / epub / mobi within given archive it returns an error as shown below.

tar: *.pdf: Not found in archive
tar: *.mobi: Not found in archive
tar: Exiting with failure status due to previous errors

How to prevent this error. I believe there should be a way to provide the multiple wildcards with a 'OR' operator as available in other scripting languages.

Answer 1

tar isn't a scripting language.

To hide the error message, just redirect the stderr of tar to a bit bucket:

tar ... 2> /dev/null

Note that you might miss other errors, though.

The safe way would be to list the files first, select the ones to extract, and only do that if there were any.

tar --list -f ...tgz | grep '\.\(pdf\|mobi\|epub\)$'

Answer 2

Thanks to @choroba below code is perfect. No error reported. Posting the code as answer so that others have better visibility to the final working piece of code.

find '/home/pi/Downloads/complete/' -type f -name "*.tgz"| while read i ; do tar --list -f "$i" | grep '\.\(pdf\|mobi\|epub\)$' | while read -r line ; do  tar -kxvzf "$i" -C "/home/pi/Downloads/complete/ebook/" "$line" --strip=1;done; done;