"David G. Johnston" <david.g.johnston@gmail.com> writes:
> Given we don't have a regexp_count function this isn't surprising...
FYI, it's there in HEAD.
In the meantime, you could possibly do something like
=# select count(*) from regexp_matches('My High Street', '([A-Z][a-z]+[\s])', 'g');
count
-------
2
(1 row)
(Note that 2 is the correct answer given that there's no space
after the third word; I trust Oracle agrees.)
regards, tom lane