Linux diff command not showing right comparison

Question

I really need your help to advice how to show the correct comparison test using diff command. I have two files

FileA:

2            1
3            4166
4            1186
5            14318
6            2451
7            64
10           3
11           18
20           5

FileB:

3            3990
4            1144
5            14368
6            2439
7            64
10           3
11           18
20           5

And running the diff command this way to show a side by side comparison, the output is not what I'm expecting:

diff -y -W 60 FileA FileB

The output is:

2            1               |  3            3992
3            4166            |  4            1144
4            1186            |  5            14372
5            14318           |  6            2439
6            2451            <
7            64                 7            64
10           3                  10           3
11           18                 11           18
20           5                  20           5

The expected output should be:

2            1               <
3            4166            |  3            3992
4            1186            |  4            1144
5            14318           |  5            14372
6            2451            |  6            2439
7            64                 7            64
10           3                  10           3
11           18                 11           18
20           5                  20           5

I mean, first line in FileB is missing, but diff is not capturing that. Can you please help me to fix this?

Answer 1

There are few ways to compare this

foo
bar

to this

baz

by comparing lines. You can say that

1. foo is only in the first file, bar and baz are the line that differs;
2. or foo and baz are the line that differs, bar is only in the first file;
3. or foo and bar are only in the first file, baz is only in the second.

Each description is true. The first two are somewhat better because in the format diff -y uses they take two lines each, while the third one takes three lines.

In this case diff -y does a good job generating two lines, not three. But still there are two possible descriptions to choose from.

You may favor the first description (where bar was replaced by baz) because this change alters just one letter; but diff has no concept of similar lines. Two lines are either different or not.

In your case, to compare 5 lines with 4 different lines, I can assume any of the 5 lines is unique to the first file and the remaining 4 lines correspond accordingly. You say the 1st line was removed; diff says the 5th one was removed. Both approaches are valid.

This is because diff works with lines and it has no concept of lines that differ only partially.

I tried wdiff, it works with words. This tool is able to identify 3, 4, 5 and 6 as identical in both your files. But it's a different tool with different output.