Merge pull request django-haystack#775 from stefanw/avoid-pks-seen-on-update

Avoid unnecessary, potentially huge db query on index update

Author: bigjust

haystack/management/commands/update_index.py

@@ -254,7 +254,6 @@ def update_backend(self, label, using):
             if self.verbosity >= 1:
                 print(u"Indexing %d %s" % (total, force_text(model._meta.verbose_name_plural)))
 
-            pks_seen = set([smart_bytes(pk) for pk in qs.values_list('pk', flat=True)])
             batch_size = self.batchsize or backend.batch_size
 
             if self.workers > 0:
@@ -278,8 +277,11 @@ def update_backend(self, label, using):
                     # They're using a reduced set, which may not incorporate
                     # all pks. Rebuild the list with everything.
                     qs = index.index_queryset().values_list('pk', flat=True)
-                    pks_seen = set([smart_bytes(pk) for pk in qs])
+                    pks_seen = set(smart_bytes(pk) for pk in qs)
+
                     total = len(pks_seen)
+                else:
+                    pks_seen = set(smart_bytes(pk) for pk in qs.values_list('pk', flat=True))
 
                 if self.workers > 0:
                     ghetto_queue = []