use bigger int for time concept

work around #124, need more consideration

ddf_utils/model/package.py

@@ -267,7 +267,7 @@ def load_ddf(self):
         # 2. dtype for time
         for c_id, c in concepts.items():
             if c.concept_type == 'time':
-                dtypes[c_id] = 'int16'  # TODO: maybe there are other format for time.
+                dtypes[c_id] = 'uint64'  # TODO: maybe there are other format for time.
         # 3. group files for same indicator together
         indicators = dict()
         for field, pkey, path in self._gen_datapoints():

tests/chef/test_procedures.py

@@ -61,7 +61,7 @@ def test_filter():
 
     assert set(dps.keys()) == {'imr_upper', 'imr_lower'}
     for dp in dps.values():
-        assert dp.year.dtype == np.int16
+        assert dp.year.dtype == np.uint64
         assert np.all(dp.year > 2000)
         assert set(dp.country.unique()) == {'usa', 'swe'}
 
@@ -81,7 +81,7 @@ def test_flatten():
         'agriculture_percentage_f', 'agriculture_percentage_m',
         'agriculture_thousands', 'agriculture_percentage'}
 
-    assert res[0].compute()['agriculture_percentage_m'].dtypes['year'] == np.int16
+    assert res[0].compute()['agriculture_percentage_m'].dtypes['year'] == np.uint64
 
 
 def test_groupby():
@@ -115,7 +115,7 @@ def test_translate_column():
     chef.run()
 
     res = chef.dag.get_node('bp-datapoints-aligned').evaluate().compute()
-    assert res['biofuels_production_kboed'].dtypes['year'] == np.int16
+    assert res['biofuels_production_kboed'].dtypes['year'] == np.uint64
 
 
 def test_translate_header():
@@ -128,7 +128,7 @@ def test_translate_header():
     assert set(data.keys()) == set(indicators)
     for i in indicators:
         assert set(data[i].columns) == set(['geo', 'time', i])
-        assert data[i].dtypes['time'] == np.int16
+        assert data[i].dtypes['time'] == np.uint64
 
     data = res[1].get_data()
     assert 'city' in data.keys()
@@ -155,7 +155,7 @@ def test_trend_bridge():
     chef.run()
 
     res = chef.dag.get_node('res-1').evaluate().compute()
-    assert res['imr_lower'].dtypes['year'] == np.int16
+    assert res['imr_lower'].dtypes['year'] == np.uint64
 
 
 def test_window():
@@ -192,7 +192,7 @@ def test_merge():
     indicators = ['imr_lower', 'imr_median', 'imr_upper',
                   'biofuels_production_kboed', 'biofuels_production_ktoe']
     assert set(data.keys()) == set(indicators)
-    assert data['imr_median'].dtypes['year'] == np.int16
+    assert data['imr_median'].dtypes['year'] == np.uint64
 
     imr_lower = data['imr_lower'].set_index(['geo', 'year'])
     assert imr_lower.loc[('afg', 1961), 'imr_lower'] == 2055