error on invalid asof_join by input (#3053)

polars/polars-core/src/frame/asof_join/groups.rs

@@ -9,6 +9,8 @@ use std::fmt::Debug;
 use std::hash::Hash;
 use std::ops::Sub;
 
+#[cfg(feature = "dtype-categorical")]
+use crate::frame::hash_join::check_categorical_src;
 use crate::frame::hash_join::{
     create_probe_table, get_hash_tbl_threaded_join_partitioned, multiple_keys as mk, prepare_strs,
 };
@@ -504,6 +506,11 @@ impl DataFrame {
                     }
                 }
             } else {
+                for (lhs, rhs) in left_by.get_columns().iter().zip(right_by.get_columns()) {
+                    check_asof_columns(lhs, rhs)?;
+                    #[cfg(feature = "dtype-categorical")]
+                    check_categorical_src(lhs.dtype(), rhs.dtype())?;
+                }
                 asof_join_by_multiple(&left_by, &right_by, left_asof, right_asof)
             }
         } else {

py-polars/tests/test_errors.py

@@ -21,3 +21,17 @@ def test_error_on_reducing_map() -> None:
         match="A 'map' functions output length must be equal to that of the input length. Consider using 'apply' in favor of 'map'.",
     ):
         df.groupby("id").agg(pl.map(["t", "y"], np.trapz))
+
+
+def test_error_on_invalid_by_in_asof_join() -> None:
+    df1 = pl.DataFrame(
+        {
+            "a": ["a", "b", "a"],
+            "b": [1, 2, 3],
+            "c": ["a", "b", "a"],
+        }
+    )
+
+    df2 = df1.with_column(pl.col("a").cast(pl.Categorical))
+    with pytest.raises(pl.ComputeError):
+        df1.join_asof(df2, on="b", by=["a", "c"])