sort: union, intersection, and ???

[vasa1]

I was reading https://github.com/jlevy/the-art-of-command-line and I found this:

      cat a b | sort | uniq > c   # c is a union b
      cat a b | sort | uniq -d > c   # c is a intersect b
      cat a b b | sort | uniq -u > c   # c is set difference a - b

Could someone please explain the third line?

[Lars Noodén]

By itself, and assuming a sorted input, uniq prints out each line once. With the -u option, only those lines that are not repeated are printed. It is the opposite of -d. So cat a b | sort | uniq -u > c would eliminate the lines that are in both set a and b, in other words a XOR b. By repeating b and using -u, it guarantees that anything in b is eliminated from the final list, whether or not it is also in a, in other words a - b.

[vasa1]

By itself, and assuming a sorted input, uniq prints out each line once. With the -u option, only those lines that are not repeated are printed. It is the opposite of -d. So cat a b | sort | uniq -u > c would eliminate the lines that are in both set a and b, in other words a XOR b. By repeating b and using -u, it guarantees that anything in b is eliminated from the final list, whether or not it is also in a, in other words a - b.

Thank you for explaining

[Lars Noodén]

No problem.

By the way, I notice that cat is unnecessary there. The utility sort can read files directly:

sort a b | uniq > c
sort a b | uniq -d > c
sort a b b | uniq -u > c

[vasa1]

No problem.

By the way, I notice that cat is unnecessary there. ...

Then it's another case of UUOC: http://porkmail.org/era/unix/award.html and http://stackoverflow.com/questions/1...ess-use-of-cat