fixedes

dios/dios.py

@@ -405,7 +405,11 @@ class DictOfSeries:
     def squeeze(self, axis=None):
         if axis in [0, 'index']:
             raise NotImplementedError
-        elif axis in [1, 'columns']:
+
+        if len(self) > 1:
+            return self
+
+        if axis in [1, 'columns']:
             return self._data.squeeze()
         elif axis is None:
             return self._data.squeeze().squeeze()

dios/locator.py

@@ -108,13 +108,14 @@ class _iLocIndexer(_Indexer):
         # we do use loc instead of iloc as we get real keys.
         # this works, because keys keep sorted if they come
         # from an series-index (doesn't work with df-index)
-        ser = self._data[key]
+        ser = self._data.loc[key]
         if ixalign:
             ix = ser.index.intersection(ix.index)
         if isinstance(right, pd.Series):
             left = ser[ix]
             right, ix = align_index_by_policy(left, right)
         ser.iloc[ix] = right
+        self._data.loc[key] = ser
 
     def _get_item(self, key, ix, ixalign=False):
         ser = self._data.loc[key]
@@ -148,7 +149,7 @@ def _unpack_value(keys, ix, val):
 
     # prepare value
     val = list(val) if is_iterator(val) else val
-    val = val.squeeze() if is_dios_like(val) else val
+    val = val.squeeze(axis=1) if is_dios_like(val) else val
     dioslike, nlistlike = is_dios_like(val), is_nested_list_like(val)
 
     # check value

profiling/performance.py

@@ -63,12 +63,12 @@ def b_timings(df, t0, t1, v1, v2):
 
 def dios_timings(dios, t0, t1, v1, v2):
     _t0 = time.time()
-    a = dios[t0:t1, :]
+    a = dios.loc[t0:t1, :]
     _t1 = time.time()
-    b = dios[:, v1]
+    b = dios.loc[:, v1]
     _t2 = time.time()
     if profile_assignment:
-        dios[t0:t1, v1] = dios[t0:t1, v1] * 1111
+        dios.loc[t0:t1, v1] = dios.loc[t0:t1, v1] * 1111
     _t3 = time.time()
 
     timingsdf.at[rows, ('ts', 'dios')] += _t1 - _t0
@@ -76,7 +76,6 @@ def dios_timings(dios, t0, t1, v1, v2):
     timingsdf.at[rows, ('ass', 'dios')] += _t3 - _t2
     return a, b, dios
 
-
 def gen_random_timestamps(m, M):
     r = (M - m) * (np.random.randint(10,90) + np.random.random()) * 0.01
     a , b = m + r, M - r
@@ -106,22 +105,22 @@ if __name__ == '__main__':
 
     # max increase of of rows
     # 1 = 10 # 2 = 100 # .... # 5 = 100'000
-    iterations = 4
-    runs = 10
+    iterations = 5
+    runs = 1
     cols = 10
 
     profile_assignment = True
 
     # which to calc and plot
-    use_df = True
-    use_a = False
-    use_b = False
+    use_df = False
+    use_a = True
+    use_b = True
     use_dios = True
 
     # plot options
-    normalize_to_df = False
+    normalize_to_df = True
     plot_xlog = True
-    plot_ylog = False
+    plot_ylog = True
 
     # ########################