feat: add tests for string equivalence in window functions and aggregations

Author: kosiew

python/tests/test_dataframe.py

@@ -933,6 +933,54 @@ def test_window_frame_defaults_match_postgres(partitioned_df):
     assert df_2.sort(col_a).to_pydict() == expected
 
 
+def test_last_value_order_by_string_equivalence(partitioned_df):
+    expr = f.last_value(column("a")).over(
+        Window(
+            partition_by=[column("c")],
+            order_by=[column("b")],
+            window_frame=WindowFrame("rows", None, None),
+        )
+    )
+    string = f.last_value(column("a")).over(
+        Window(
+            partition_by=[column("c")],
+            order_by="b",
+            window_frame=WindowFrame("rows", None, None),
+        )
+    )
+    df = partitioned_df.select(expr.alias("expr"), string.alias("str"))
+    table = pa.Table.from_batches(df.collect())
+    assert table.column("expr").to_pylist() == table.column("str").to_pylist()
+
+
+def test_nth_value_order_by_string_equivalence(partitioned_df):
+    expr = f.nth_value(column("b"), 3).over(Window(order_by=[column("a")]))
+    string = f.nth_value(column("b"), 3).over(Window(order_by="a"))
+    df = partitioned_df.select(expr.alias("expr"), string.alias("str"))
+    table = pa.Table.from_batches(df.collect())
+    assert table.column("expr").to_pylist() == table.column("str").to_pylist()
+
+
+def test_rank_order_by_string_equivalence(partitioned_df):
+    expr = f.rank(order_by=[column("b")])
+    string = f.rank(order_by="b")
+    df = partitioned_df.select(expr.alias("expr"), string.alias("str"))
+    table = pa.Table.from_batches(df.collect())
+    assert table.column("expr").to_pylist() == table.column("str").to_pylist()
+
+
+def test_array_agg_order_by_string_equivalence(partitioned_df):
+    df = partitioned_df.aggregate(
+        [column("c")],
+        [
+            f.array_agg(column("a"), order_by=[column("a")]).alias("expr"),
+            f.array_agg(column("a"), order_by="a").alias("str"),
+        ],
+    ).sort(column("c"))
+    table = pa.Table.from_batches(df.collect())
+    assert table.column("expr").to_pylist() == table.column("str").to_pylist()
+
+
 def test_html_formatter_cell_dimension(df, clean_formatter_state):
     """Test configuring the HTML formatter with different options."""
     # Configure with custom settings