Comparing directory files based on filename

Question

I currently have a directory of files that looks something like this:

abcd.txt
abcd_.txt
qrst.txt
qrst_.txt
wxyz.txt
wxyz_.txt

In theory, every line in abcde_.txt should be contained in abcd.txt, and every line in qrst_.txt should be contained in qrst.txt, and so on. While I have no problem comparing two files to test for this individually, I'm trying to find a more efficient way to do this for the entire directory.

In a case like this, if I had a lot of pairs of files, but I didn't know in advance what string of letters they'd each start with, is there a way to loop through and process each set of two related files at a time?

Answer 1

From the list of name, you could find *_.txt. From the name, remove _.txt and add .txt. Compare the two files.

for f1 in *_.txt; do
    f2="${f1%_.txt}.txt"
    compare "$f1" "$f2"
done

Answer 2

If you already know that for every file file.txt, there is a file file_.txt, I will use the _ to differentiate between them

A code like this should work:

files=(`ls | grep -v "_"`)
for f1 in ${files[@]}
do
     f2=`echo $f1 | sed -r 's/.txt/_.txt/g'`
     diff $f1 $f2
done