Excise EOF character from AWK processing - awk

Question

Watch this:
echo one two three | awk 'BEGIN {RS=" "; FS=""} {print NF}'
3
3
6
That last count is suspicious. Let's have a look at the last field of each record:
echo one two three | awk 'BEGIN {RS=" "; FS=""} {print $NF}'
e
o
e
I will take a wild guess and assume AWK is picking up some sort of EOF character after the last field of the last record.
How do I tell AWK to ignore such extraneous character?

Answer 1

The input stream includes a trailing \n.
By default awk will treat that trailing \n as the record separator (as opposed to part of the data).
In your code you've redefined awk's record separator as a space, thus leaving the \n to be treated as part of the input data.
If your version of awk supports multiple record separators you could define a dual record separator of space and \n, eg:
$ echo one two three | awk 'BEGIN {RS="[ \n]"; FS=""} {print NF}'
3
3
5
Or, as mentioned in the comment, you could explicitly remove the \n, eg:
$ echo one two three | awk 'BEGIN {RS=" "; FS=""} {sub(/\n/,"");print NF}'
3
3
5
Alternatively you could strip the \n before calling awk, eg:
$ echo one two three | tr -d '\n' | awk 'BEGIN {RS=" "; FS=""} {print NF}'
3
3
5
Or you could even skip the generation of the \n in the first place, eg:
$ printf "one two three" | awk 'BEGIN {RS=" "; FS=""} {print NF}'
3
3
5