The statement generated by utils.dictToWhere is not suitable for None value

[Dmitri Bogomolov]

The statement should be IS NULL for None value, not = NULL. As a result BDObject.findBy() returns [] in that case. The following test succeeds if name is a string:

diff --git a/twistar/tests/test_dbobject.py b/twistar/tests/test_dbobject.py
index 6611e90..d7001d9 100644
--- a/twistar/tests/test_dbobject.py
+++ b/twistar/tests/test_dbobject.py
@@ -84,6 +84,12 @@ class DBObjectTest(unittest.TestCase):
         resultids = [result.id for result in results]
         self.assertEqual(ids, resultids)

+    @inlineCallbacks
+    def test_null(self):
+        name = None
+        user = yield User(first_name=name).save()
+        users = yield User.findBy(first_name=name)
+        self.assertEqual(user, users[0])

     @inlineCallbacks
     def test_count(self):

[Dmitri Bogomolov]

I think a solution may be like this

diff --git a/twistar/utils.py b/twistar/utils.py
index 91e7ee1..8b5c1f1 100644
--- a/twistar/utils.py
+++ b/twistar/utils.py
@@ -67,7 +67,8 @@ def dictToWhere(attrs, joiner="AND"):
     if len(attrs) == 0:
         return None

-    wheres = map(lambda name: "(%s = ?)" % name, attrs.keys())
+    wheres = ["(%s = ?)" % key if val else "(%s IS ?)" % key
+              for key, val in attrs.iteritems()]
     return [(" %s " % joiner).join(wheres)] + attrs.values()

[Brian Muller]

Great catch! Thanks!