full order by support with arrange; closes #30

siuba/sql/translate.py

@@ -2,6 +2,7 @@
 from sqlalchemy.sql import sqltypes as types
 from functools import singledispatch
 from .verbs import case_when, if_else
+import warnings
 
 # TODO: must make these take both tbl, col as args, since hard to find window funcs
 def sa_is_window(clause):
@@ -44,6 +45,13 @@ class CumlOver(Over, CustomOverClause):
     def set_over(self, group_by, order_by):
         self.partition_by = group_by
         self.order_by = order_by
+
+        if not len(order_by):
+            warnings.warn(
+                    "No order by columns explicitly set in window function. SQL engine"
+                    "does not guarantee a row ordering. Recommend using an arrange beforehand.",
+                    RuntimeWarning
+                    )
         return self
 
 

siuba/sql/verbs.py

@@ -382,9 +382,10 @@ def _arrange(__data, *args):
             for expr in args
             )
 
-    sort_cols = _create_order_by_clause(cols, *args)
+    sort_cols = _create_order_by_clause(cols, *new_calls)
 
-    return __data.append_op(last_op.order_by(*sort_cols), order_by = new_calls)
+    order_by = __data.order_by + new_calls
+    return __data.append_op(last_op.order_by(*sort_cols), order_by = order_by)
 
 
 # TODO: consolidate / pull expr handling funcs into own file?
@@ -492,8 +493,7 @@ def _summarize(__data, **kwargs):
         sel.append_column(col)
 
     # TODO: is a simple method on __data for doing this...
-    new_data = __data.append_op(sel)
-    new_data.group_by = tuple()
+    new_data = __data.append_op(sel, group_by = tuple(), order_by = tuple())
     return new_data
 
 

siuba/tests/test_verb_filter.py

@@ -5,7 +5,7 @@
 """
 
 from siuba import _, filter, group_by, arrange
-from siuba.dply.vector import row_number
+from siuba.dply.vector import row_number, desc
 import pandas as pd
 
 import pytest
@@ -61,3 +61,13 @@ def test_filter_via_group_by_arrange(backend):
             data_frame(g = [1, 2, 2], x = [3, 3, 4])
             )
 
+@backend_notimpl("sqlite")
+def test_filter_via_group_by_desc_arrange(backend):
+    dfs = backend.load_df(x = [3,2,1] + [2,3,4], g = [1]*3 + [2]*3)
+
+    assert_equal_query(
+            dfs,
+            group_by(_.g) >> arrange(desc(_.x)) >> filter(_.x.cumsum() > 3),
+            data_frame(g = [1, 1, 2, 2, 2], x = [2, 1, 4, 3, 2])
+            )
+

siuba/tests/test_verb_mutate.py

@@ -81,11 +81,14 @@ def test_mutate_using_agg_expr(backend):
 def test_mutate_using_cuml_agg(backend):
     data = data_frame(x = range(1, 5), g = [1,1,2,2])
     dfs = backend.load_df(data)
-    assert_equal_query(
-            dfs,
-            group_by(_.g) >> mutate(y = _.x.cumsum()),
-            data.assign(y = [1.0, 3, 3, 7])
-            )
+
+    # cuml window without arrange before generates warning
+    with pytest.warns(None):
+        assert_equal_query(
+                dfs,
+                group_by(_.g) >> mutate(y = _.x.cumsum()),
+                data.assign(y = [1.0, 3, 3, 7])
+                )
 
 @pytest.mark.skip("TODO: mutate not preserving var order (#42)")
 def test_mutate_overwrites_prev(backend):

siuba/tests/test_verb_summarize.py

@@ -66,7 +66,6 @@ def test_summarize_no_same_call_var_refs(df):
         df >> summarize(y = _.x.min(), z = _.y + 1)
 
 
-@pytest.mark.skip("TODO (see #30)")
 def test_summarize_removes_order_vars(df):
     lazy_tbl = df >> summarize(n = n(_))