comm question

webeye · 11-24-2005, 09:56 PM

hello
i use comm to compare two very simple file.
file1:
0
0
1

file2:
0
1
0

comm file1 file2 ---->output:
................0
0
................1
........0

I view comm's man page, the interpretation of output is listed following:

Column one contains lines unique to FILE1, column two contains lines unique to FILE2, and column three contains lines common to both files

But i do not understand how does the result generate above? it seems different from the man page gave. Anyone can tell me what's the rule of comm's comparison or give me a reference? Thank in advance

zahadumy · 11-25-2005, 08:23 PM

That's very weird. I used comm a couple of times and it worked fine for me. I suppose it's a bug and that happens when you have the same line more than one time in a file:

Code:

[cosmin ~]$ cat file1
0
1
2
3
[cosmin ~]$ cat file2
2
3
4
5
[cosmin ~]$ comm file1 file2
0
1
                2
                3
        4
        5

And now this is what I suppose it's a bug:

Code:

[cosmin ~]$ cat file1
0
1
2
3
1
2
[cosmin ~]$ cat file2
2
3
4
5
1
2
[cosmin ~]$ comm file1 file2
0
1
                2
                3
1
2
        4
        5
        1
        2

My opinion is that you have nothing to understand here. If I were you, I would write a C program which does this.

eddiebaby1023 · 11-26-2005, 11:02 AM

Sigh. If you read the man page for comm you might notice the words "compare two sorted files line by line". Note the word "sorted". The bug is in your use of comm, not in the code itself.

zahadumy · 11-26-2005, 11:07 AM

Good point. Thank you eddiebaby1023.

eddiebaby1023 · 11-27-2005, 01:11 PM

Quote:

Originally posted by zahadumy
Good point. Thank you eddiebaby1023.

I got bitten the first time I used comm.

webeye · 11-27-2005, 09:39 PM

First, thank you for zahadumy and eddiebaby1023 very much.

Second, I see that the two file must be sorted. but i don't understand what does the mean of "line by line"?

etc:
file1
0
1

file2
1
2

comm output
0
...................1
...........2

Who can describe the output of comparable process of simple example above?

zahadumy · 11-28-2005, 04:06 AM

Quote:

Originally posted by webeye
First, thank you for zahadumy and eddiebaby1023 very much.

Second, I see that the two file must be sorted. but i don't understand what does the mean of "line by line"?

etc:
file1
0
1

file2
1
2

comm output
0
...................1
...........2

Who can describe the output of comparable process of simple example above?

It compares the file line by line, as you can see. On the first column you see lines which appear only in the first file, on the second column lines which appear only in the second file and on the third one lines common to both files. For more details:

Code:

man comm
info comm

"info comm" displays the complete manual

bigearsbilly · 11-28-2005, 10:41 AM

they must be sorted first.

Quote:

The comm utility will read file1 and file2, which should be
ordered in the current collating sequence, and produce three
text columns as output: lines only in file1; lines only in

same goes for 'uniq'

webeye · 11-28-2005, 07:57 PM

`comm' writes to standard output lines that are common, and lines that are unique

above is the first line of "info comm". Is the part of text bold means the line of file also must unique?

etc:
0
1
-----input is right
0
0
1
-----input is wrong, because the 0 is not unique

is right about my understand above?

zahadumy · 11-29-2005, 12:49 AM

Quote:

Originally posted by webeye
`comm' writes to standard output lines that are common, and lines that are unique

above is the first line of "info comm". Is the part of text bold means the line of file also must unique?

etc:
0
1
-----input is right
0
0
1
-----input is wrong, because the 0 is not unique

is right about my understand above?

No, you can give as input whatever you want, but the file must be sorted.