Enable doc tests in local and CI testing (#1409)

* Turn on doctests

* Fix existing doc examples

* Remove stale referenece to rust-toolchain removed in #1383, surpised pre-commit didn't flag for anyone else

Author: ntjohnson1

.github/workflows/test.yml

@@ -62,7 +62,7 @@ jobs:
         uses: actions/cache@v5
         with:
           path: ~/.cargo
-          key: cargo-cache-${{ steps.rust-toolchain.outputs.cachekey }}-${{ hashFiles('Cargo.lock') }}
+          key: cargo-cache-${{ matrix.toolchain }}-${{ hashFiles('Cargo.lock') }}
 
       - name: Install dependencies
         uses: astral-sh/setup-uv@v7
@@ -106,7 +106,7 @@ jobs:
           RUST_BACKTRACE: 1
         run: |
           git submodule update --init
-          uv run --no-project pytest -v . --import-mode=importlib
+          uv run --no-project pytest -v --import-mode=importlib
 
       - name: FFI unit tests
         run: |

pyproject.toml

@@ -70,6 +70,9 @@ features = ["substrait"]
 [tool.pytest.ini_options]
 asyncio_mode = "auto"
 asyncio_default_fixture_loop_scope = "function"
+addopts = "--doctest-modules"
+doctest_optionflags = ["NORMALIZE_WHITESPACE", "ELLIPSIS"]
+testpaths = ["python/tests", "python/datafusion"]
 
 # Enable docstring linting using the google style guide
 [tool.ruff.lint]

python/datafusion/dataframe.py

@@ -327,8 +327,9 @@ def into_view(self, temporary: bool = False) -> Table:
             >>> df = ctx.sql("SELECT 1 AS value")
             >>> view = df.into_view()
             >>> ctx.register_table("values_view", view)
-            >>> df.collect()  # The DataFrame is still usable
-            >>> ctx.sql("SELECT value FROM values_view").collect()
+            >>> result = ctx.sql("SELECT value FROM values_view").collect()
+            >>> result[0].column("value").to_pylist()
+            [1]
         """
         from datafusion.catalog import Table as _Table
 
@@ -1389,9 +1390,12 @@ def fill_null(self, value: Any, subset: list[str] | None = None) -> DataFrame:
             DataFrame with null values replaced where type casting is possible
 
         Examples:
-            >>> df = df.fill_null(0)  # Fill all nulls with 0 where possible
-            >>> # Fill nulls in specific string columns
-            >>> df = df.fill_null("missing", subset=["name", "category"])
+            >>> from datafusion import SessionContext, col
+            >>> ctx = SessionContext()
+            >>> df = ctx.from_pydict({"a": [1, None, 3], "b": [None, 5, 6]})
+            >>> filled = df.fill_null(0)
+            >>> filled.sort(col("a")).collect()[0].column("a").to_pylist()
+            [0, 1, 3]
 
         Notes:
             - Only fills nulls in columns where the value can be cast to the column type