Package: moreutils
Version: 0.18
Severity: wishlist

Hi,

I often have the need to do operations like that:
  Get the lines from file1, ie:
    pkg1 ver1 comment1
    pkg2 ver2 comment2
    pkg3 ver3 comment3
    pkg4 ver4 comment4
  Where the first field is not in file2, ie:
    pkg2
    pkg3

combine file1 not file2 obviously won't work, since the whole lines are
compared. It would be great if one could specify a subset of fields on
which to compare, using cut's syntax for example, so I could write:
combine -d ' ' -f 1 file1 not file2

Of course, that causes problem with duplicate hash keys, and the cases
where the rest of the lines differ between the files. I propose the
following policy:
  for and/not:
    the lines in file1 are outputted
  for or/xor:
    the line in file1 is outputted, then the line if file2 if no
    matching line in file1

Additionnally, 'not' could be managed as a special case, not 'unifying'
the lines in file1, that is:
    pkg1 ver1 comment1
    pkg1 ver11 comment11
    pkg2 ver2 comment2
    pkg2 ver22 comment22
    pkg3 ver3 comment3
    pkg3 ver33 comment33
    pkg4 ver4 comment4
    pkg4 ver44 comment44
  not
    pkg2
    pkg3
would output:
    pkg1 ver1 comment1
    pkg1 ver11 comment11
    pkg4 ver4 comment4
    pkg4 ver44 comment44

Anyway, all this stuff should probably be implemented together with some
test cases.

This is an old wishlist item, but I think it would make the combine utility
much better.

I suggest using the same parameter names as join for specifying fields to
compare.

       -t CHAR
              use CHAR as input and output field separator

       -1 FIELD
              join on this FIELD of file 1

       -2 FIELD
              join on this FIELD of file 2