Expression iterator

packages/vaex-core/vaex/expression.py

@@ -531,6 +531,21 @@ def to_pandas_series(self):
         import pandas as pd
         return pd.Series(self.values)
 
+    def __len__(self):
+        return len(self.df)
+
+    def __iter__(self):
+        n_samples = len(self)
+        chunk_size = 1000
+
+        def iterator():
+            for ndx in range(0, n_samples, chunk_size):
+                i1, i2 = ndx, min(ndx + chunk_size, n_samples)
+                values = self[i1:i2].values.tolist()
+                for value in values:
+                    yield value
+        return iterator()
+
     def __getitem__(self, slicer):
         """Provides row and optional field access (struct arrays) via bracket notation.
 

tests/getattr_test.py

@@ -70,6 +70,18 @@ def test_expression(ds_local):
     assert np.array(dss).T.tolist() == [(ds.y/10).values.tolist(), (ds.x/5).values.tolist()]
 
 
+def test_expression_iterator(ds_local):
+    ds = ds_local
+
+    assert len(ds) == len(ds.x)
+    assert len(ds[ds.x > 1]) == len(ds[ds.x > 1].x)
+
+    ds = ds_local.head(10)
+    for (x_value, y_value), x_iter_value, y_iter_value in zip(ds[['x', 'y']].values, ds.x, ds.y):
+        assert x_value == x_iter_value
+        assert y_value == y_iter_value
+
+
 @pytest.mark.skip(reason="Not implemented yet, should work, might need refactoring of copy")
 def test_expression_virtual(ds_local):
     ds = ds_local