I have a short test file (named "dict") comprising 25 words, each on a separate line. A few of these words are the plural form of the word above (e.g. "aardvark" followed by "aardvarks"). I want to mark all the words that end with an "s" as part of a larger project that will eventually delete plural forms from a much larger file (over 100,000 words), but my only concern at this stage of my self-education is to target those words that end with an "s" by replacing the terminal "s" with an "x".

I thought it could be done with the following:
However, that returns exactly the contents of the "dict" file with no replacements.

I can replace all "s" characters with
and I can replace all initial "a" characters with
but the end-of-line $ isn't doing what I expected.

I'm using Kubuntu 12.04, if that makes any difference. Any ideas where I should look for a fix?

Perhaps you should investigate the obvious first - maybe the "s" isn't the last character in the line. Whitespace(s) can be difficult to discern.

2 members found this post helpful.

syg00 -- That helped! I did a hex dump of the "dict" file and found that it had apparently been created in an MS Windows environment. There was a "0d 0a" at the end of each line instead of simply a "0a". I got rid of all those superfluous carriage-return codes with
After that, the original sed code worked fine!
whizje, the code you suggested changes the entire word (after I modified the file to delete the carriage-return codes) to a single "x", not just the trailing "s" characters. A variation on that may be useful as I develop my sed script further, but it's not exactly what I was aiming at right now.

For converting line endings, see the tool http://linux.die.net/man/1/dos2unix and unix2dos.
Usually part of the std install and also handles MAC line endings as well.