Skip to content

Commit 8ab223f

Browse files
committed
chore: update Changelog
1 parent c173b81 commit 8ab223f

1 file changed

Lines changed: 61 additions & 1 deletion

File tree

‎CHANGELOG.md‎

Lines changed: 61 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -5,7 +5,67 @@
55

66
### Features
77

8-
* add polars support ([f8ba5ef](https://github.com/getyourguide/dataframe-expectations/commit/f8ba5ef6932e84d83081da6971552ea972bab062))
8+
* **Polars support** ([f8ba5ef](https://github.com/getyourguide/dataframe-expectations/commit/f8ba5ef6932e84d83081da6971552ea972bab062))
9+
10+
`dataframe-expectations` now supports [Polars](https://pola.rs/) DataFrames alongside Pandas and PySpark. Polars is an optional dependency and must be explicitly requested:
11+
12+
```
13+
pip install dataframe-expectations[polars] # pandas + polars
14+
pip install dataframe-expectations[pyspark,polars] # pandas + pyspark + polars
15+
```
16+
17+
18+
All existing expectations that have a Polars implementation will automatically dispatch to the
19+
correct validation path when a Polars DataFrame is passed to a suite.
20+
21+
Polars imports are fully deferred behind `@lru_cache` helpers (mirroring the PySpark pattern),
22+
so importing `dataframe_expectations` has zero overhead when Polars is not installed.
23+
24+
* **Backward-compatible API for expectation authors** ([52dda74](https://github.com/getyourguide/dataframe-expectations/commit/52dda741fa5e3376dd4072a1d8dc1ece33ec8248))
25+
26+
The `fn_violations_polars` parameter on `ColumnExpectation` is optional (`None` by default),
27+
so existing custom expectations that only implement Pandas and/or PySpark continue to work
28+
without changes. `@abstractmethod` has been removed from `validate_pandas`, `validate_pyspark`,
29+
and `validate_polars` — each raises `NotImplementedError` by default, allowing expectation
30+
authors to implement only the backends they need.
31+
32+
* **CI updated to cover Polars install scenarios** ([b636a15](https://github.com/getyourguide/dataframe-expectations/commit/b636a151b6aa9864723e9b11551533c0139d2538))
33+
34+
| Job | Extras installed | Tests run |
35+
|---|---|---|
36+
| `tests-without-optional` | None | `-m "not pyspark and not polars"` |
37+
| `tests-with-pyspark` | `[pyspark]` | `-m "not polars"` |
38+
| `tests-with-polars` | `[polars]` | `-m "not pyspark"` |
39+
| `tests-with-all` | `[pyspark,polars]` | All |
40+
41+
42+
### Bug Fixes
43+
44+
* **Tightened DataFrame type checks** ([15d1df7](https://github.com/getyourguide/dataframe-expectations/commit/15d1df7a3e5b0d1aa397a562fef8ac5ae1c49e6b))
45+
46+
`is_polars_data_frame()` and `is_pyspark_data_frame()` no longer use `__module__` name
47+
fallbacks. Only actual `DataFrame` instances (via `isinstance`) are accepted, preventing
48+
`Series`, `Expr`, `LazyFrame`, `Column`, `Row`, etc. from being mis-routed into validation paths.
49+
50+
51+
### Documentation
52+
53+
* **Getting Started guide updated with tabbed examples** ([ee53581](https://github.com/getyourguide/dataframe-expectations/commit/ee53581e1b8fe50ed296a476e43f010ca25e865a))
54+
55+
Installation and usage examples now use `sphinx-design` tab sets showing Pandas, PySpark, and
56+
Polars side-by-side. `sphinx-design>=0.6.0` added as a docs dependency.
57+
58+
59+
### Test Improvements
60+
61+
* **Refactored test fixtures to use PyArrow** ([f9692bd](https://github.com/getyourguide/dataframe-expectations/commit/f9692bd))
62+
63+
Test data is now defined once via PyArrow tables and converted to Pandas, PySpark, or Polars
64+
DataFrames through a `dataframe_factory` fixture, eliminating duplicated test data across
65+
backends. All expectation tests are parametrized over `[pandas, pyspark, polars]`.
66+
67+
* **Exhaustive `match/case` blocks** — all `match df_lib` blocks in tests now include
68+
`case _: pytest.fail(...)` to catch unhandled DataFrame types immediately.
969

1070
## [0.6.0](https://github.com/getyourguide/dataframe-expectations/compare/v0.5.2...v0.6.0) (2026-03-18)
1171

0 commit comments

Comments
 (0)