I have two files with contents of some words.
Example
file1 it has 10 words:
world
earth
eye
ear
near
from
going
want
we
our
Second file2 has 7 words:
world
earth
eye
ear
near
going
want
I want output as third file for the words which does not exist in file2.
for example (from , our , we) does not exist in file2.
Just do
grep -vFxf file2 file1 > file3
This will return those lines that doesn't exist in file2 but in file1 and write the result to file3.
-v, reverse match, here means those lines if exist only in file2. Without -v it will return those lines which are exist in both files.
-F, this is telling grep to match the pattern as a fixed pattern string instead of regex (regular expressions)
-x, matches the whole line as pattern string
-f, reading the patterns from a fileOr per your question's title and the command you referred to that sort -u, seems you want those unique words (actually lines) either exist in the file1 or file2. then you just need.
uniq -u <(sort file1 file2) > file3
sort file1 file2 | uniq -u. Note that using -w here only does what you want/say if there's only one word per line (and no non-word character), so you might as well use -x instead of -w.
Another solution:
comm -23 <(sort file1) <(sort file2) > file3
The -23 will suppress the rows contained in either only file2 or in both files.
