How do I grep, or using awk or sed, a file to find and list and expression part of which stays the same and part of which changes. Also, I would like to not have the whole line listed, just the words found?
for example, in the file are many different instances of 'ABC=123456.' ABC stays the same, but 123456 chasnges, both in the number of integers and the integers themselves. And, when I display the results of the grep, I would like it to display each line with ABC=123456 only, not the whole line. If I use grep -o ABC=, I just get ABC=.

Thanks

Try man grep or info grep.

grep alone probably won't do that.
grep for the lines you want, e.g. "grep ABC= file"
then pipe into a sed command to replace what you don't want with nothing, i.e. "s/ABC=//"
Cheers, Tom.

If you could show an actual example (input & desired output) it would be better than just trying to describe it.

Here is an example. There are many lines like the one below in the log file. I have made many substitutions in the example below because it is proprietary information.ABC=XXXXXXXX is all that I want from the lines. ABC is a constant; those aren't the letters but there are three letters that stay the same.What ABC equals in each line is different in each line (it is not XXXXs). And the rest of the information in this line is even more than what is in this example, and it varies from line to line.

Thanks

0150310
ERROR 2012-04-10 15:03:37,556 : , ABC=XXXXXX Title=XXXXXXXXX XXXXXXX: XXXXXXXX XXXXXXX =sysdate Date=20120410150337

what I would like it to look like:

ABC=XXXXXXX Title=XXXXXXXXXX Date=XXXXXXXXXXXXX

Make up your mind.
The response in #2 should give you all you need to construct a solution.

Yes it will.....
grep works fine for the problem as first stated (try the code given earlier)

Note: the "search term" with grep is a regular expression (Regex) ((or extended Regex, if you use egrep)).

or