single-row vs array update

gistfile1.sql

-- row by row
for c1 in (select rowid rid, YEAR1,
                    QUARTER,
                    ...
      from flights_400k
	  )
  loop
-- moderately complex PL/SQL logic 
update flights_400k
   set DEPARTUREDELAYGROUPS = l_group,
	   DEPDELAYMINUTES 		= 10
 where rowid = c1.rid;
 commit;
end loop;
  
-- array update
open flights_cur;
loop
	fetch flights_cur
		bulk collect into 
		l_rowid,l_group,l_delaymin limit l_limit;

	exit when l_rowid.count = 0;
	
	for i in l_rowid.first..l_rowid.last
	loop
	-- do your PL/SQL logic here
		l_group(i) := 1;
		l_delaymin(i) := 10;
	end loop; --i
  
	-- 1 update of the table in bulk
	forall j in indices of l_rowid
	update /*+ parallel(flights_400k) */ flights_400k
	set DEPARTUREDELAYGROUPS 	= l_group(j),
		DEPDELAYMINUTES 		= l_delaymin(j)
	where rowid = l_rowid(j);
	
	dbms_output.put_line('Modified ' || SQL%ROWCOUNT || ' rows');
	
end loop;