Remove DISTINCT from node_query_node_access_alter()

Well not exactly, merely move the distinct to a sub query. A DISTINCT clause means the database server has to iterate over the returned result set and insure a single column has no to exact same values. That becomes a large task when the result set is large. However, if we added the DISTINCT clause as a subquery to the table that produces the duplicate entries (node_access), then the DISTINCT filter would filter over the least amount of results needed to produce a unique result set.

Alternately, the sub-query DISTINCT could become a "COUNT(*), nid" and GROUP BY nid which will also produce unique nid records.Though since grant_(op) is a required field, this is a less desirable tactic.