grep command not finding text in MS-Office or PDF documents

Question

Debian GNU/Linux 11 (bullseye), grep (GNU grep) 3.6

I need find string in current directory within all files (doc, docx and pdf), grep command not working for me:

grep -ril "word" .

It doesn't output anything. What's wrong?

Answer 1

All three formats need to be converted to text before they can be searched using tools such as grep.

For "old-style" .doc files, use catdoc:

catdoc file.doc | grep word

For OOXML .docx files, use docx2txt:

docx2txt < file.docx | grep word

or

docx2txt file.docx - | grep word

For PDF files, use pdfgrep:

pdfgrep word file.pdf

or pdftotext:

pdftotext file.pdf - | grep word

If you switch to ripgrep you can use a preprocessor:

#!/bin/sh -
if [ ! -s "1ドル" ]; then exec cat; fi
case "1ドル" in
*.pdf)
 exec pdftotext - -
 ;;
*.doc)
 exec catdoc -
 ;;
*.docx)
 exec docx2txt - -
 ;;
*)
 exec cat
 ;;
esac

Save this to a file, make it executable (chmod 755), and use it with --pre:

rg --pre /path/to/preprocessor word

See the ripgrep guide for tips on reducing the overhead of the preprocessor.