UNIX for Dummies Questions & Answers

Greetings.

I like to compare two lists of numbers, A.txt and B.txt, to see the numbers that are in B.txt but not in A.txt. I only need the "deletions" with reference to A.txt. Using the diff command doesn't work because it considers too many of the deletions just changes because they are a digit different than one in A.txt.

Input,

A.txt:

Code:

0113429
0113430
0113510
0113590
0113993

B.txt:

Code:

0113429
0113430
0113510
0113589
0113590

Output.txt:

Code:

added
0113993

deleted
0113589

I would like something that only accepts or reject perfect matchs, changes are not of interest.

Thanks!

Show the input you have, and the output you want.

Done, thanks!

diff works perfectly for this data, so it cannot be an example of your problem.

What would some data that doesn't work with diff look like? What digits are allowed to differ and still be considered the 'same'?

The data there was just an illustration, the actual files are lists of 1501 and 451 numbers respectively. The difference in number of lines is likely another issue.

The match must be 100% perfect to be considered non-unique and discarded. The output should only be of unique elements, any change would be considered a new and non-unique element.

The comm utility sounds like what you want, then. It outputs three columns -- lines only found in file 1, lines only found in file 2, and lines found in both. It needs sorted (or at least identical) order but since you were using diff, it sounds like we can depend on that. You can suppress any of the columns with the -1, -2, -3 options.

Code:

comm -2 -3 file1 file2 > onlyinfile1
comm -1 -3 file1 file2 > onlyinfile2

Using grep command with -f and -v option:

Code:

echo -e "deleted\n$( grep -f A.txt -v B.txt )"
echo -e "added\n$( grep -f B.txt -v A.txt )"