I have two text files to compare: file A and  file B

file A is nicely formatted with sections, headings etc, for better visibility

file B is a linear raw list of strings (really a bunch of machine names)

I am trying to compare file A and file B, and locate the strings in each file that don't exist in the other, and vice versa-- in other words, identify unique strings in each file.

UNIX utility *diff* works great, so do Windows tools like 'ExamDiff', 'CompareIt!' etc; but they only compare a single occurence of each string pair, and ignore the rests.

For instance, I have

List A        List B
------       ------
abc          bcd
def          def
def          ijk    
ghi           jkl

The result will be:

List A        List B
------        ------
abc           bcd
def           ijk
ghi            jkl

(Note that the eliminated strings were the ones that followed One-to-One matching)

While the expected result is:

List A        List B
------      ------
abc         bcd
ghi          ijk
              jkl

With both occurences of 'def' being eliminated - with One-to-many comparison.

Can anyone suggest a solution? A tool or an script logic?

~@~

try http://www.scootersoftware.com/download.php

Just installed and tried it -- still the same issue ..it does the comparison, but only for a single occurrence. I haven't had a chance to look at the options yet though.

Thanks!

~@~

Download the 30 day free trail of arexis merge. This tool works great.
http://www.araxis.com/merge/

Have you tried this tool for similar purpose?
Will take a look. Thanks!

BTW, I was able to get it done on Excel (underrated, but worked great) by playing around with logical comparison formulas.

~@~

Create two text files.
x.txt with
abc
def
def
ghi

and y.txt with
bcd
def
ijk
jkl

Issue following commands (at cygwin prompt)
//sort and copy unique elements of x to x1
sort -u x.txt > x1.txt
//sort and copy unique elements of y to y1
sort -u y.txt > y1.txt

//copy lines that appear in the both x1 and y1 to z
comm -1 -2 x1.txt y1.txt >z.txt

//output lines that appear in x1 only
comm -2 -3 x1.txt z.txt

//output lines that appear in y1 only
comm -2 -3 y1.txt z.txt

Works for this limited dataset.
Give it a try.

Thanks @gidilat, for the advice, and welcome to the forum - if you are a new member :-)

Just looked through the commands and spotted a minor gotcha..

When you do sort -u and sort the list in order by unique elements, you'd actually get rid of multiple occurrences of each element. Once that's done, it's really a one to one comparison, no?

In the above scenario, ALL occurrences of def on list A were eliminated - as they matched def on list B.
But if I had, say abc listed twice in list A and did not exist at all in list B, sort -u would remove the second instance of abc in list A, correct? But the goal is to have every single occurrence of abc in list A if it's not present in list B.

Thanks again for chiming in.
~@~

You are changing the goal from
"in other words, identify unique strings in each file"
to
"goal is to have every single occurrence of abc in list A if it's not present in list'

A problem needs to defined properly before it can be solved.

Hmm, so - I said:
"...the goal is to have every single occurrence of abc in list A if it's not present in list B"

Doesn't that leave each list's strings/elements unique to those of the other list (and hence "...unique strings in each file")?

When we compare two files and speak about 'uniqueness', I'd think the reference would be toward uniqueness with respect to each other, and not within oneself.

Regardless, I'll buy the fact that 'unique strings in each file with respect to each other' or something similar would've made it little more descriptive. Sorry -- Thanks!

So, just for the heck of it - I tried comm without sorting and isolating the unique elements, works just as good as any other alternatives (except Excel) that I tried in the past. It does the comparison, but only gets rid of a single occurrence of the match.

~@~