test and better error messages

pola-rs · Sep 18, 2024 · d3cb162 · d3cb162
1 parent 6208079
commit d3cb162
Show file tree

Hide file tree

Showing 2 changed files with 21 additions and 2 deletions.
diff --git a/crates/polars-core/src/datatypes/dtype.rs b/crates/polars-core/src/datatypes/dtype.rs
@@ -793,10 +793,11 @@ pub fn merge_dtypes(left: &DataType, right: &DataType) -> PolarsResult<DataType>
             let merged = merge_dtypes(inner_l, inner_r)?;
             List(Box::new(merged))
         },
+        #[cfg(feature = "dtype-struct")]
         (Struct(inner_l), Struct(inner_r)) => {
-            polars_ensure!(inner_l.len() == inner_r.len(), ComputeError: "cannot combine different structs");
+            polars_ensure!(inner_l.len() == inner_r.len(), ComputeError: "cannot combine structs with differing amounts of fields ({} != {})", inner_l.len(), inner_r.len());
             let fields = inner_l.iter().zip(inner_r.iter()).map(|(l, r)| {
-                polars_ensure!(l.name() == r.name(), ComputeError: "cannot combine different structs");
+                polars_ensure!(l.name() == r.name(), ComputeError: "cannot combine structs with different fields ({} != {})", l.name(), r.name());
                 let merged = merge_dtypes(l.dtype(), r.dtype())?;
                 Ok(Field::new(l.name().clone(), merged))
             }).collect::<PolarsResult<Vec<_>>>()?;

diff --git a/py-polars/tests/unit/datatypes/test_categorical.py b/py-polars/tests/unit/datatypes/test_categorical.py
@@ -7,6 +7,8 @@
 
 import pytest
 
+from hypothesis import given
+
 import polars as pl
 from polars import StringCache
 from polars.exceptions import (
@@ -845,3 +847,19 @@ def test_get_cat_categories_multiple_chunks() -> None:
     )
     df_cat = df.lazy().select(pl.col("e").cat.get_categories()).collect()
     assert len(df_cat) == 2
+
+
+@pytest.mark.parametrize(
+    "f", [
+        lambda x: (pl.List(pl.Categorical), [x]),
+        lambda x: (pl.Struct({ 'a': pl.Categorical }), { 'a': x }),
+    ]
+)
+def test_nested_categorical_concat(f: Callable[[str], tuple[pl.DataType, list[str] | dict[str, str]]]) -> None:
+    dt, va = f("a")
+    _, vb = f("b")
+    a = pl.DataFrame({ 'x': [va] }, schema={ 'x': dt })
+    b = pl.DataFrame({ 'x': [vb] }, schema={ 'x': dt })
+
+    with pytest.raises(pl.exceptions.StringCacheMismatchError):
+        stack = pl.concat([a, b])