diff --git a/dataflows/VERSION b/dataflows/VERSION
index 31c01c9..1df5b7e 100644
--- a/dataflows/VERSION
+++ b/dataflows/VERSION
@@ -1 +1 @@
-0.0.45
+0.0.46
diff --git a/dataflows/base/datastream_processor.py b/dataflows/base/datastream_processor.py
index 28af4d3..d5525a0 100644
--- a/dataflows/base/datastream_processor.py
+++ b/dataflows/base/datastream_processor.py
@@ -11,6 +11,15 @@
 from .schema_validator import schema_validator
 
 
+class LazyIterator:
+
+    def __init__(self, get_iterator):
+        self.get_iterator = get_iterator
+
+    def __iter__(self):
+        return self.get_iterator()
+
+
 class DataStreamProcessor:
 
     def __init__(self):
@@ -38,31 +47,37 @@ def process_row(self, row):
     def process_datapackage(self, dp: Package):
         return dp
 
+    def get_res(self, current_dp, name):
+        ret = self.datapackage.get_resource(name)
+        if ret is None:
+            ret = current_dp.get_resource(name)
+        assert ret is not None
+        return ret
+
+    def get_iterator(self, datastream):
+        current_dp = datastream.dp
+        res_iter_ = datastream.res_iter
+
+        def func():
+            res_iter = (ResourceWrapper(self.get_res(current_dp, rw.res.name), rw.it)
+                        for rw in res_iter_)
+            res_iter = self.process_resources(res_iter)
+            res_iter = (it if isinstance(it, ResourceWrapper) else ResourceWrapper(res, it)
+                        for res, it
+                        in itertools.zip_longest(self.datapackage.resources, res_iter))
+            return res_iter
+        return func
+
     def _process(self):
         datastream = self.source._process()
-        current_dp = datastream.dp
 
         self.datapackage = Package(descriptor=copy.deepcopy(datastream.dp.descriptor))
         self.datapackage = self.process_datapackage(self.datapackage)
         self.datapackage.commit()
 
-        res_iter = datastream.res_iter
-
-        def get_res(name):
-            ret = self.datapackage.get_resource(name)
-            if ret is None:
-                ret = current_dp.get_resource(name)
-            assert ret is not None
-            return ret
-
-        res_iter = (ResourceWrapper(get_res(rw.res.name), rw.it)
-                    for rw in res_iter)
-        res_iter = self.process_resources(res_iter)
-        res_iter = (it if isinstance(it, ResourceWrapper) else ResourceWrapper(res, it)
-                    for res, it
-                    in itertools.zip_longest(self.datapackage.resources, res_iter))
-
-        return DataStream(self.datapackage, res_iter, datastream.stats + [self.stats])
+        return DataStream(self.datapackage,
+                          LazyIterator(self.get_iterator(datastream)),
+                          datastream.stats + [self.stats])
 
     def process(self):
         ds = self._process()
diff --git a/dataflows/helpers/iterable_loader.py b/dataflows/helpers/iterable_loader.py
index ad6d9cf..570ab9a 100644
--- a/dataflows/helpers/iterable_loader.py
+++ b/dataflows/helpers/iterable_loader.py
@@ -77,7 +77,7 @@ def handle_iterable(self):
         try:
             for x in self.iterable:
                 if mode is None:
-                    assert isinstance(x, (dict, list))
+                    assert isinstance(x, (dict, list)), 'Bad item %r' % x
                     mode = dict if isinstance(x, dict) else list
                 assert isinstance(x, mode)
                 if mode == dict:
diff --git a/dataflows/processors/delete_fields.py b/dataflows/processors/delete_fields.py
index 0927979..2461408 100644
--- a/dataflows/processors/delete_fields.py
+++ b/dataflows/processors/delete_fields.py
@@ -3,9 +3,11 @@
 
 def process_resource(rows, fields):
     for row in rows:
-        for field in fields:
-            del row[field]
-        yield row
+        yield dict(
+            (k, v)
+            for k, v in row.items()
+            if k not in fields
+        )
 
 
 def delete_fields(fields, resources=None):
diff --git a/dataflows/processors/printer.py b/dataflows/processors/printer.py
index e9a9cda..749a7b4 100644
--- a/dataflows/processors/printer.py
+++ b/dataflows/processors/printer.py
@@ -59,10 +59,10 @@ def func(rows):
         x = 1
 
         for i, row in enumerate(rows):
-            yield row
 
             index = i + 1
-            row = [index] + [truncate_cell(row[f], max_cell_size) for f in field_names]
+            prow = [index] + [truncate_cell(row[f], max_cell_size) for f in field_names]
+            yield row
 
             if index - x == (num_rows + 1):
                 x *= num_rows
@@ -71,9 +71,9 @@ def func(rows):
                 last.clear()
                 if toprint and toprint[-1][0] != index - 1:
                     toprint.append(['...'])
-                toprint.append(row)
+                toprint.append(prow)
             else:
-                last.append(row)
+                last.append(prow)
                 if len(last) > (last_rows or num_rows):
                     last = last[1:]