What is the most efficient way to find the record count of a large table?

Here are the two way's i've did it, would just like opinions on which one is better and more efficient. I'm coding PHP which shouldn't make a difference I'm just wondering about the SQL syntax.

or
in the above after selecting all I use the mysql_num_rows function.

Randy

The first returns only a single row (and column)

The second returns the entire data from the table.

Now think about which is more efficient...

That's what I figured but I also figured it had to go through the whole table to get the count.

no, it doesn't always do that -- myisam tables store the count, so getting the count is lightning fast

but even if the database has to count all the rows, that's still a lot faster than exporting all the rows to let you count them