ADD PRINTS

.github/workflows/test-python.yml

@@ -90,6 +90,14 @@ jobs:
           python tests/docs/run_doctest.py
           pytest tests/docs/test_user_guide.py -m docs
 
+      - name: Run Multiscan Parametric
+        if: github.ref_name != 'main'
+        run: pytest -n auto --dist=loadgroup -m "not release and not benchmark and not docs" tests/unit/io/test_multiscan.py -k test_multiscan_slice_parametric -s
+
+      - name: Run Multiscan Slice
+        if: github.ref_name != 'main'
+        run: pytest -n auto --dist=loadgroup -m "not release and not benchmark and not docs" tests/unit/io/test_multiscan.py -k test_multiscan_slice_middle -s
+
       - name: Run tests
         if: github.ref_name != 'main'
         run: pytest -n auto --dist=loadgroup -m "not release and not benchmark and not docs" tests/unit/io/test_multiscan.py -s

py-polars/tests/unit/io/test_multiscan.py

@@ -26,6 +26,8 @@
     ],
 )
 def test_include_file_paths(tmp_path: Path, scan: Any, write: Any) -> None:
+    print(f"START: {tmp_path}")
+
     a_path = tmp_path / "a"
     b_path = tmp_path / "b"
 
@@ -44,6 +46,8 @@ def test_include_file_paths(tmp_path: Path, scan: Any, write: Any) -> None:
         ),
     )
 
+    print(f"END: {tmp_path}")
+
 
 @pytest.mark.parametrize(
     ("scan", "write", "ext", "supports_missing_columns", "supports_hive_partitioning"),
@@ -72,7 +76,8 @@ def test_multiscan_projection(
     hive: bool,
     col: bool,
 ) -> None:
-    print(tmp_path)
+    print(f"START: {tmp_path}")
+
     a = pl.DataFrame({"col": [5, 10, 1996]})
     b = pl.DataFrame({"col": [13, 37]})
 
@@ -156,6 +161,8 @@ def test_multiscan_projection(
                 .collect(new_streaming=True),  # type: ignore[call-overload]
             )
 
+    print(f"END: {tmp_path}")
+
 
 @pytest.mark.parametrize(
     ("scan", "write", "ext"),
@@ -172,6 +179,8 @@ def test_multiscan_row_index(
     write: Callable[[pl.DataFrame, Path], Any],
     ext: str,
 ) -> None:
+    print(f"START: {tmp_path}")
+
     a = pl.DataFrame({"col": [5, 10, 1996]})
     b = pl.DataFrame({"col": [42]})
     c = pl.DataFrame({"col": [13, 37]})
@@ -228,6 +237,8 @@ def test_multiscan_row_index(
         ),
     )
 
+    print(f"END: {tmp_path}")
+
 
 @pytest.mark.parametrize(
     ("scan", "write", "ext"),
@@ -257,6 +268,8 @@ def test_schema_mismatch_type_mismatch(
     write: Callable[[pl.DataFrame, Path], Any],
     ext: str,
 ) -> None:
+    print(f"START: {tmp_path}")
+
     a = pl.DataFrame({"xyz_col": [5, 10, 1996]})
     b = pl.DataFrame({"xyz_col": ["a", "b", "c"]})
 
@@ -276,6 +289,8 @@ def test_schema_mismatch_type_mismatch(
     ):
         q.collect(new_streaming=True)  # type: ignore[call-overload]
 
+    print(f"END: {tmp_path}")
+
 
 @pytest.mark.parametrize(
     ("scan", "write", "ext"),
@@ -305,6 +320,8 @@ def test_schema_mismatch_order_mismatch(
     write: Callable[[pl.DataFrame, Path], Any],
     ext: str,
 ) -> None:
+    print(f"START: {tmp_path}")
+
     a = pl.DataFrame({"x": [5, 10, 1996], "y": ["a", "b", "c"]})
     b = pl.DataFrame({"y": ["x", "y"], "x": [1, 2]})
 
@@ -321,6 +338,8 @@ def test_schema_mismatch_order_mismatch(
     with pytest.raises(pl.exceptions.SchemaError):
         q.collect(new_streaming=True)  # type: ignore[call-overload]
 
+    print(f"END: {tmp_path}")
+
 
 @pytest.mark.parametrize(
     ("scan", "write"),
@@ -342,6 +361,8 @@ def test_multiscan_head(
     scan: Callable[..., pl.LazyFrame],
     write: Callable[[pl.DataFrame, io.BytesIO | Path], Any],
 ) -> None:
+    print(f"START: HEAD-{scan}")
+
     a = io.BytesIO()
     b = io.BytesIO()
     for f in [a, b]:
@@ -353,6 +374,8 @@ def test_multiscan_head(
         pl.Series("c1", range(5)).to_frame(),
     )
 
+    print(f"END: HEAD-{scan}")
+
 
 @pytest.mark.parametrize(
     ("scan", "write"),
@@ -374,6 +397,8 @@ def test_multiscan_tail(
     scan: Callable[..., pl.LazyFrame],
     write: Callable[[pl.DataFrame, io.BytesIO | Path], Any],
 ) -> None:
+    print(f"STARt: TIAL-{scan}")
+
     a = io.BytesIO()
     b = io.BytesIO()
     for f in [a, b]:
@@ -385,6 +410,8 @@ def test_multiscan_tail(
         pl.Series("c1", range(5, 10)).to_frame(),
     )
 
+    print(f"END: TAIL-{scan}")
+
 
 @pytest.mark.parametrize(
     ("scan", "write"),
@@ -406,6 +433,7 @@ def test_multiscan_slice_middle(
     scan: Callable[..., pl.LazyFrame],
     write: Callable[[pl.DataFrame, io.BytesIO | Path], Any],
 ) -> None:
+    print(f"START: SLICE-{scan}")
     fs = [io.BytesIO() for _ in range(13)]
     for f in fs:
         write(pl.Series("c1", range(7)).to_frame(), f)
@@ -441,6 +469,7 @@ def test_multiscan_slice_middle(
         scan(fs, row_index_name="ri").slice(offset, 17).collect(new_streaming=True),  # type: ignore[call-overload]
         pl.DataFrame(ri_expected_series),
     )
+    print(f"END: SLICE-{scan}")
 
 
 @pytest.mark.parametrize(
@@ -471,6 +500,8 @@ def test_multiscan_slice_parametric(
     offset: int,
     length: int,
 ) -> None:
+    print(f"START: SLICE-{ext}-{offset}-{length}")
+
     # Once CSV negative slicing is implemented this should be removed. If we
     # don't do this, this test is flaky.
     if ext == "csv":
@@ -509,3 +540,5 @@ def test_multiscan_slice_parametric(
         .slice(offset, length)
         .collect(new_streaming=True),  # type: ignore[call-overload]
     )
+
+    print(f"END: SLICE-{ext}-{offset}-{length}")