diff --git a/strax/context.py b/strax/context.py
index 3fe7c475..2f58f262 100644
--- a/strax/context.py
+++ b/strax/context.py
@@ -1908,15 +1908,8 @@ def get_df(
 
         """
         df = self.get_array(run_id, targets, save=save, max_workers=max_workers, **kwargs)
-        try:
-            return pd.DataFrame.from_records(df)
-        except Exception as e:
-            if "Data must be 1-dimensional" in str(e):
-                raise ValueError(
-                    f"Cannot load '{targets}' as a dataframe because it has "
-                    "array fields. Please use get_array."
-                )
-            raise
+
+        return strax.convert_structured_array_to_df(df, log=self.log)
 
     def get_zarr(
         self,
diff --git a/strax/utils.py b/strax/utils.py
index 3f3ea41b..430c0e48 100644
--- a/strax/utils.py
+++ b/strax/utils.py
@@ -804,3 +804,40 @@ def convert_tuple_to_list(init_func_input):
     else:
         # if not a container, return. i.e. int, float, bytes, str etc.
         return func_input
+
+
+@export
+def convert_structured_array_to_df(structured_array, log=None):
+    """Convert a structured numpy array to a pandas DataFrame.
+
+    Parameters:
+    structured_array (numpy.ndarray): The structured array to be converted.
+    Returns:
+    pandas.DataFrame: The converted DataFrame.
+
+    """
+
+    if log is None:
+        import logging
+
+        log = logging.getLogger("strax_array_to_df")
+
+    data_dict = {}
+    converted_cols = []
+    for name in structured_array.dtype.names:
+        col = structured_array[name]
+        if col.ndim > 1:
+            # Convert n-dimensional columns to lists of ndarrays
+            data_dict[name] = [np.array(row) for row in col]
+            converted_cols.append(name)
+        else:
+            data_dict[name] = col
+
+    if converted_cols:
+        log.warning(
+            f"Columns {converted_cols} contain non-scalar entries. "
+            "Some pandas functions (e.g., groupby, apply) might "
+            "not perform as expected on these columns."
+        )
+
+    return pd.DataFrame(data_dict)
diff --git a/tests/test_core.py b/tests/test_core.py
index ca3063eb..b821b29e 100644
--- a/tests/test_core.py
+++ b/tests/test_core.py
@@ -33,6 +33,28 @@ def test_core(allow_multiprocess, max_workers, processor):
     assert bla.dtype == strax.peak_dtype()
 
 
+@processing_conditions
+def test_core_df(allow_multiprocess, max_workers, processor, caplog):
+    """Test that get_df works with N-dimensional data."""
+    """Test that get_df works with N-dimensional data."""
+    mystrax = strax.Context(
+        storage=[],
+        register=[Records, Peaks],
+        processors=[processor],
+        allow_multiprocess=allow_multiprocess,
+        use_per_run_defaults=True,
+    )
+
+    df = mystrax.get_df(run_id=run_id, targets="peaks", max_workers=max_workers)
+    p = mystrax.get_single_plugin(run_id, "records")
+    assert len(df.loc[0, "data"]) == 200
+    assert len(df) == p.config["recs_per_chunk"] * p.config["n_chunks"]
+    assert (
+        "contain non-scalar entries. Some pandas functions (e.g., groupby, apply)"
+        " might not perform as expected on these columns." in caplog.text
+    )
+
+
 def test_post_office_state():
     mystrax = strax.Context(
         storage=[],