ovn-northd-ddlog: Optimize AggregatedFlow rules.

This should avoid some work by doing the cheapest check (the one on
UseLogicalDatapathGroups) before any joins.  DDlog is probably
factoring out the reference to the Flow relation, which is identical
in both, but this ought to avoid the group_by aggregation (which is
relatively expensive) in the case where UseLogicalDatapathGroups is
not enabled.

Acked-by: Han Zhou <hzhou@ovn.org>
Acked-by: Dumitru Ceara <dceara@redhat.com>
Signed-off-by: Ben Pfaff <blp@ovn.org>
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: Mark Michelson <mmichels@redhat.com>

northd/ovn_northd.dl

@@ -1689,18 +1689,18 @@ AggregatedFlow(.logical_datapaths = g.to_set(),
                .actions = actions,
                .tags = tags,
                .external_ids = external_ids) :-
+    UseLogicalDatapathGroups[true],
     TaggedFlow(logical_datapath, stage, priority, __match, actions, tags, external_ids),
-    var g = logical_datapath.group_by((stage, priority, __match, actions, tags, external_ids)),
-    UseLogicalDatapathGroups[true].
+    var g = logical_datapath.group_by((stage, priority, __match, actions, tags, external_ids)).
 AggregatedFlow(.logical_datapaths = set_singleton(logical_datapath),
                .stage = stage,
                .priority = priority,
                .__match = __match,
                .actions = actions,
                .tags = tags,
                .external_ids = external_ids) :-
-    TaggedFlow(logical_datapath, stage, priority, __match, actions, tags, external_ids),
-    UseLogicalDatapathGroups[false].
+    UseLogicalDatapathGroups[false],
+    TaggedFlow(logical_datapath, stage, priority, __match, actions, tags, external_ids).
 
 for (f in AggregatedFlow()) {
     var pipeline = if (f.stage.pipeline == Ingress) "ingress" else "egress" in