test: Add tests for various open issues (#20720)

py-polars/tests/unit/constructors/test_constructors.py

@@ -1182,6 +1182,31 @@ def test_from_dicts_schema_columns_do_not_match() -> None:
     assert_frame_equal(result, expected)
 
 
+def test_from_dicts_infer_integer_types() -> None:
+    data = [
+        {
+            "a": 2**7 - 1,
+            "b": 2**15 - 1,
+            "c": 2**31 - 1,
+            "d": 2**63 - 1,
+            "e": 2**127 - 1,
+        }
+    ]
+    result = pl.from_dicts(data).schema
+    # all values inferred as i64 except for values too large for i64
+    expected = {
+        "a": pl.Int64,
+        "b": pl.Int64,
+        "c": pl.Int64,
+        "d": pl.Int64,
+        "e": pl.Int128,
+    }
+    assert result == expected
+
+    with pytest.raises(OverflowError):
+        pl.from_dicts([{"too_big": 2**127}])
+
+
 def test_from_rows_dtype() -> None:
     # 50 is the default inference length
     # 5182

py-polars/tests/unit/datatypes/test_decimal.py

@@ -606,3 +606,18 @@ def test_decimal_horizontal_20482() -> None:
         "max": [D("123.000000"), D("234.000000")],
         "sum": [D("246.000000"), D("468.000000")],
     }
+
+
+def test_shift_over_12957() -> None:
+    df = pl.DataFrame(
+        {
+            "a": [1, 1, 2, 2],
+            "b": [D("1.1"), D("1.1"), D("2.2"), D("2.2")],
+        }
+    )
+    result = df.select(
+        x=pl.col("b").shift(1).over("a"),
+        y=pl.col("a").shift(1).over("b"),
+    )
+    assert result["x"].to_list() == [None, D("1.1"), None, D("2.2")]
+    assert result["y"].to_list() == [None, 1, None, 2]

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

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+import io
 import tempfile
 from collections import OrderedDict
 from pathlib import Path
@@ -438,3 +439,11 @@ def test_scan_csv_with_column_names_nonexistent_file() -> None:
     # Upon collection, it should fail
     with pytest.raises(FileNotFoundError):
         result.collect()
+
+
+def test_select_nonexistent_column() -> None:
+    csv = "a\n1"
+    f = io.StringIO(csv)
+
+    with pytest.raises(pl.exceptions.ColumnNotFoundError):
+        pl.scan_csv(f).select("b").collect()

py-polars/tests/unit/streaming/test_streaming_join.py

@@ -308,3 +308,16 @@ def test_flush_join_and_operation_19040() -> None:
         "B": [None, 1],
         "C": [None, 1],
     }
+
+
+def test_full_coalesce_join_and_rename_15583() -> None:
+    df1 = pl.LazyFrame({"a": [1, 2, 3]})
+    df2 = pl.LazyFrame({"a": [3, 4, 5]})
+
+    result = (
+        df1.join(df2, on="a", how="full", coalesce=True)
+        .select(pl.all().name.map(lambda c: c.upper()))
+        .sort("A")
+        .collect(streaming=True)
+    )
+    assert result["A"].to_list() == [1, 2, 3, 4, 5]

py-polars/tests/unit/test_empty.py

@@ -158,3 +158,10 @@ def test_empty_input_expansion() -> None:
                 pl.col("B").sort_by(pl.struct(pl.exclude("A", "B")))
             )
         )
+
+
+def test_empty_list_15523() -> None:
+    s = pl.Series("", [["a"], []], dtype=pl.List)
+    assert s.dtype == pl.List(pl.String)
+    s = pl.Series("", [[], ["a"]], dtype=pl.List)
+    assert s.dtype == pl.List(pl.String)