default negative offset in dynamic groupby (#2823)

py-polars/Makefile

@@ -11,8 +11,8 @@ venv:
 	@unset CONDA_PREFIX && source venv/bin/activate && maturin develop
 
 clean:
-	@rm -r venv
-	@rm -r target/
+	@-rm -r venv
+	@cargo clean
 
 pre-commit: venv
 	$(PYTHON_BIN)/isort .

py-polars/polars/internals/frame.py

@@ -2610,7 +2610,7 @@ def groupby_dynamic(
         period
             length of the window, if None it is equal to 'every'
         offset
-            offset of the window
+            offset of the window if None and period is None it will be equal to negative `every`
         truncate
             truncate the time value to the window lower bound
         include_boundaries

py-polars/polars/internals/lazy_frame.py

@@ -834,7 +834,7 @@ def groupby_dynamic(
         period
             length of the window, if None it is equal to 'every'
         offset
-            offset of the window
+            offset of the window if None and period is None it will be equal to negative `every`
         truncate
             truncate the time value to the window lower bound
         include_boundaries
@@ -848,10 +848,13 @@ def groupby_dynamic(
 
         """
 
+        if offset is None:
+            if period is None:
+                offset = f"-{every}"
+            else:
+                offset = "0ns"
         if period is None:
             period = every
-        if offset is None:
-            offset = "0ns"
         by = _prepare_groupby_inputs(by)
         lgb = self._ldf.groupby_dynamic(
             index_column,

py-polars/tests/test_datelike.py

@@ -539,3 +539,29 @@ def test_epoch() -> None:
     assert dates.dt.epoch("d").series_equal(
         (dates.dt.timestamp("ms") // (1000 * 3600 * 24)).cast(pl.Int32)
     )
+
+
+def test_default_negative_every_offset_dynamic_groupby() -> None:
+    # 2791
+    dts = [
+        datetime(2020, 1, 1),
+        datetime(2020, 1, 2),
+        datetime(2020, 2, 1),
+        datetime(2020, 3, 1),
+    ]
+    df = pl.DataFrame({"dt": dts, "idx": range(len(dts))})
+    out = df.groupby_dynamic(index_column="dt", every="1mo", closed="right").agg(
+        pl.col("idx")
+    )
+
+    expected = pl.DataFrame(
+        {
+            "dt": [
+                datetime(2020, 1, 1, 0, 0),
+                datetime(2020, 1, 1, 0, 0),
+                datetime(2020, 3, 1, 0, 0),
+            ],
+            "idx": [[0], [1, 2], [3]],
+        }
+    )
+    assert out.frame_equal(expected)