BUG: handle empty lists in json_normalize

xref: pandas-dev#15534

Author: rgbkrk

pandas/io/json/normalize.py

@@ -157,6 +157,9 @@ def _pull_field(js, spec):
 
         return result
 
+    if isinstance(data, list) and len(data) is 0:
+        return DataFrame()
+
     # A bit of a hackjob
     if isinstance(data, dict):
         data = [data]

pandas/tests/io/json/test_normalize.py

@@ -62,6 +62,11 @@ def test_simple_normalize(self):
 
         tm.assert_frame_equal(result, expected)
 
+    def test_empty_array(self):
+        result = json_normalize([])
+        expected = DataFrame()
+        tm.assert_frame_equal(result, expected)
+
     def test_more_deeply_nested(self):
         data = [{'country': 'USA',
                  'states': [{'name': 'California',