Good day everyone,

Thank you in advance for any assistance you can provide. Here is my boggle. I'm working on a Centos 6.5 server where I'm pulling an active directory group with winbind. I'm using wbinfo to obtain the listing of the domain admins group to check if there is a difference so I can add those new admins to approximately 300 servers. I am using Chef to push it out through a recipe.

O.k., there is a short history. I'm using the following command to pull the info from AD:

wbinfo --group-info "domain admins" > sysadmin.txt

Which gives me the following output:

domain admins:*:12345678:me,you,them

I then copy that file to sysadmin_old.txt and then run the wbinfo command again and compare the new file with the old file. So, the two files look like this:

sysadmin_old.txt
domain admins:*:12345678:me,you,them

sysadmin.txt
domain admins:*:12345678:me,you,them,us

I am trying to get the true difference between these two files; "us" into a 3rd file named diff.txt.

So, the diff.txt file will look like this:

us

I have used the following with no success:

grep -F -x sysadmin_old.txt sysadmin.txt > diff.txt
grep -f sysadmin_old.txt sysadmin.txt
diff -u sysadmin_old.txt sysadmin.txt | grep -E "^\+"
comm -1 sysadmin_old.txt sysadmin.txt
comm -2 sysadmin_old.txt sysadmin.txt
comm -3 sysadmin_old.txt sysadmin.txt
comm -13 sysadmin_old.txt sysadmin.txt
grep -Fxvf sysadmin_old.txt sysadmin.txt > diff.txt
diff sysadmin_old.txt sysadmin.txt | grep "<" | sed 's/^<//g' > diff.txt
sort sysadmin_old.txt sysadmin.txt | uniq -u > diff.txt
diff sysadmin_old.txt sysadmin.txt | perl -lne 'print $1 if(/^\> (.*)/)'

I have spent the last several hours googling this issue and have been unsuccessful. Any help you can provide is greatly appreciated. I will admit I am just now learning sed and awk, so there may be a solution there that I have not found.

Thank you!

Sincerely,
lhiggie1

I'm not very proficient with awk or sed either, but perhaps dwdiff will do what you want.
Using the following command:
yields this output:
You could pipe this to awk with "FS="+" to obtain ",us" and then pipe to "tr" to get rid of the comma(s).

HTH

3 members found this post helpful.

Norobro,

Thank you!!!! I'm not familiar with dwdiff. Installed it and it worked perfectly!

I'll mark as solved and add this to my bag of goodies.

Sincerely,
lhiggie1