How do I Count the words in a string using regex

I’m trying to count the words in a string using regex in Oracle 10g. I’ve been trying this

select *
from books
where REGEXP_LIKE(title, '[ ]{2}'); 

so that its returning titles with at least 3 words in the title.

Answer

INSTR is also a viable option. By looking for the second occurrence of a space, that will indicate that the string has at least 3 words.

WITH
    books
    AS
        (SELECT 'Tom Sawyer' title FROM DUAL
         UNION ALL
         SELECT 'A tale of two cities' FROM DUAL
         UNION ALL
         SELECT 'The Little Prince' FROM DUAL
         UNION ALL
         SELECT 'Don Quixote' FROM DUAL)
SELECT title
  FROM books
 WHERE instr(title, ' ', 1, 2) > 0;

If you do with to stick with regex, the regex expression below can be used to find books that have 3 or more words.

WITH
    books
    AS
        (SELECT 'Tom Sawyer' title FROM DUAL
         UNION ALL
         SELECT 'A tale of two cities' FROM DUAL
         UNION ALL
         SELECT 'The Little Prince' FROM DUAL
         UNION ALL
         SELECT 'Don Quixote' FROM DUAL)
SELECT title
  FROM books
 WHERE REGEXP_LIKE (title, '(S+s){2,}');

(Thanks @Littlefoot for the books!)