In SELECT(): Use a more efficient algorithm for keeping track

In SELECT(): Use a more efficient algorithm for keeping track
of the "aggregate" tag.  Instead of making an extra pass using
jsonObjectFindPath(), look for the tag as we examine each
field in the SELECT list.  Build a GROUP BY list every time,
but only use it later if "aggregate" was found (or if the
DISTINCT flag is set).

git-svn-id: svn://svn.open-ils.org/ILS/trunk@12302 dcc99617-32d9-48b4-a31d-7c20da2025e4