How to print lines in one file that do not match lines in another after transformation

Question

Please note, i understand how to output lines in one file that are not in another (here), my question is a little different.

In one file i have lines akin to

Андреев
Барбашев
Иванов
...

in a different file there are lines:

Барбашёв
Семёнов
...

Now. I need the lines from the second file, but only if you cannot find a line in the first where you substitute ё for е. For example Барбашёв should not display, because Барбашев is in the first.

If i do something like

comm -13 first.txt <(cat second.txt | sed 's/ё/е/g')

i get the correct lines, however, they have already been tranformed by that time, and it's unacceptable for what i'm trying to do.

In other words the output is:

Барбашев
...

While it should be

Барбашёв
...

What do you mean they are already transformed, please post full expected output. — 123, Dec 02 '16 at 09:27
@123 Sorry, i thought it was clear enough. But i've added the expected output. — v010dya, Dec 02 '16 at 09:29
`comm` works on sorted files. Have you tried sorting both files before using `comm`? You should not change anything else. — GMichael, Dec 02 '16 at 09:29
what about `grep -vf file1 file2`? This uses `file1` to obtain the patterns. — fedorqui, Dec 02 '16 at 09:33
@123 That's not going to match all the time then, only on words with no `е` in them. — v010dya, Dec 02 '16 at 09:35
@v010dya I really don't think i understand what you want to do. — 123, Dec 02 '16 at 09:44
@123 It's ok. I think it might be not doable with regular tools, i'll write a 10 line C++ program i think, it'll be faster. — v010dya, Dec 02 '16 at 09:51

FloHe · Accepted Answer · 2016-12-02T11:45:59.270

You meant:

"Now. I need the lines from the second file, but only if you cannot find a line in the first when you substitute ё for е in the second file."

instead of

"Now. I need the lines from the second file, but only if you cannot find a line in the first where you substitute ё for е."

Right?

Without using a cyrilic charset, this solution works:

file test.awk

#!/usr/bin/gawk -f

{
    if(NR==FNR)
        arr[$1]++;
    else {

        tmp=$1;
        gsub("t","e",tmp)

        if(!(tmp in arr))
            printf("%s\n", $1);
    }
}

Use:

$ ./test.awk file1 file2

If you substitute "t" -> "ё" this should also work imo. Maybe you can try.

How to print lines in one file that do not match lines in another *after transformation*

1 Answers1

How to print lines in one file that do not match lines in another after transformation