shawn wrote: > > shawn wrote: > > > comm -3 file1 file2 > > > Print lines in file1 not in file2, and vice versa. > > > seemingly conflicts with the documentation give above: > > > -3 suppress column 3 (lines that appear in both files) > > If you supressed column 3, wouldn't you end up with all > non-common/unique lines in BOTH files, according to the lower > documentation.
Yes. In column one and column two. The comm documentation says: Column one contains lines unique to FILE1, column two contains lines unique to FILE2, and column three contains lines common to both files. Suppressing column three would leave column one and column two. Column one contains lines unique to FILE1. Column two contains lines unique to FILE2. Or said another way, it prints lines in file1 not in file2, and vice versa. I think the main point of your objection is that while "print" is correct that it doesn't specify the column it is printed in. The language is less precise than in the behavior description which it more explicitly describes what is happening. However this is in an example which is extra additional information. I think it is good for examples to use alternate language. Many people find it useful to have alternate descriptions since different people read things differently. Simply repeating the previous behavior description verbatim would be less useful to a large number of people even if it is the most precise. > Also, needs a warning in man page about needing to sort files first, > which was the real problem I was having. (found doc here: > http://pubs.opengroup.org/onlinepubs/009604599/utilities/comm.html ) Well... The info documentation is the primary documentation for all of the GNU software. In the comm documentation it says: Before `comm' can be used, the input files must be sorted using the collating sequence specified by the `LC_COLLATE' locale. The man page is generated but in the man page it says: Compare sorted files FILE1 and FILE2 line by line. And there is also the --check-order option too. So there is already pretty strong wording in all of the documentation that the input must be sorted. Additionally the comm program itself will generate a warning if it finds that the input is not sorted. Example: $ comm <(printf "one\ntwo\nthree") <(printf "three\ntwo\none\n") one three two comm: file 1 is not in sorted order comm: file 2 is not in sorted order one three In the end there is only so much that can be done. Bob
signature.asc
Description: Digital signature
