Can you please let us know the regex variant you are using? While they generally perform the same operation, there are functional and syntactic differences between them all.
I'm making a few assumptions (such as "<arbitrary number (even 0) of chars,digits or special chars>" is another way of saying "0 or more non-whitespace characters") and it would be nice to have some sample text to work with. However try this:
^\S*\s+[\d.]+\s+[a-z]+\S*$
with the "ignore case" option set. If you are presenting a block of text and expect the regex to process each line, then you will also need the "multiline" option set.
I have not tested this pattern but it should be close.
Susan