Well, if it's just one directory, you don't need find, as suggested by others. A for-loop and file globbing should suffice. For the editing task, ed will do nicely.

Code:

for fname in example*; do
    [ -f "$fname" ] || continue
    printf '%s\n' 'g/user@example\.com/s//user2@example.com/g' w q | ed -s "$fname"
done

If you are familiar with the nuances of the here-document, you may prefer this more readable version:

Code:

for fname in example*; do
    [ -f "$fname" ] || continue
    ed -s "$fname" <<-'EOED'
         g/user@example\.com/s//user2@example.com/g
         w
         q
    EOED
done

Note that no other characters (invisible or not) may occur on the line where "EOED" terminates the here-document, except leading tabs (not spaces).

Regards,
Alister

---------- Post updated at 12:14 PM ---------- Previous update was at 11:57 AM ----------

There shouldn't be any pitfalls looping with a POSIX-compliant sh:

Code:

find ... | while IFS= read -r filename; do ...

Both IFS= and -r are necessary to read text verbatim. The former prevents leading and trailing whitespace from being discarded; the latter disables backslash escaping.

The only limitation that remains is the inability to read a filename with an embedded newline. In my opinion, such a filename is pathological and should be rejected with extreme prejudice.

Regards,
Alister