@@ -79,4 +79,7 @@ Miscellaneous improvements
 Fixes
 ^^^^^
 
-* fixed something (closes :issue:`1`).
+* fixed loading data files with all the following conditions when using
+  Pandas >= 3.0: mixed type columns, some combination of labels in index
+  columns not being present in the data and a non-NaN fill_value
+  (closes :issue:`1166`).
@@ -88,6 +88,29 @@ def cartesian_product_df(df, sort_rows=False, sort_columns=False,
             np.array_equal(idx.values, prod_index.values) and
             np.array_equal(columns.values, prod_columns.values)):
         return df, combined_labels
+    import numbers
+    if (isinstance(fill_value, numbers.Number) and not np.isnan(fill_value) and
+        any(dt.kind == 'O' and dt.type is str for dt in df.dtypes)):
+        df = df.copy()
+        for col in df.columns:
+            dt = df[col].dtype
+            if dt.kind == 'O' and dt.type is str:
+                # TODO: we should really output this warning, but the user
+                #       needs a way to silence it, which requires we implement
+                #       in all user-facing functions calling this function
+                #       directly or indirectly (from_series, from_frame,
+                #       read_excel, read_csv, ...) a way to specify
+                #       both a usecols argument (when the data has mixed type
+                #       but the user only needs an homogeneously typed subset)
+                #       and a dtype=object argument when the user does need
+                #       mixed types (and want to silence the warning).
+                # warnings.warn("fill_value is not valid for all "
+                #               "columns because it is a (non-NaN) number but "
+                #               f"the '{col}' column has string dtype. That "
+                #               "column will converted to object dtype to avoid "
+                #               "errors but this may cause performance issues.",
+                #               FutureWarning, stacklevel=3)
+                df[col] = df[col].astype(object)
     return df.reindex(index=prod_index, columns=prod_columns,
                       fill_value=fill_value, **kwargs), combined_labels
 
 
@@ -4656,6 +4656,38 @@ def test_from_frame():
     res = from_frame(df, sort_rows=True, sort_columns=True)
     assert_larray_equal(res, expected)
 
+    # h) with a MultiIndex with some missing combination and a fill_value which
+    #    is not valid for all columns. This pattern broke for Pandas >= 3
+    #    See issue #1166. In the user code, this issue came from, from_frame
+    #    is used indirectly by read_excel and users are keeping only the
+    #    numeric column (col2 in the test below) just after reading the array.
+    df = pd.DataFrame({
+        'idx1': [0, 0, 1],
+        'idx2': ['a', 'b', 'a'],
+        'col1': ['a', 'a', 'a'],
+        'col2': [1, 1, 1]
+    }).set_index(['idx1', 'idx2'])
+    # TODO: we should have this warning, but before that we need more
+    #       functionality to let users do something about it.
+    #       See comment in inout/pandas.py:cartesian_product_df
+    # msg = ("fill_value is not valid for all columns because it is a "
+    #        "(non-NaN) number but the 'col1' column has string dtype. That "
+    #        "column will converted to object dtype to avoid errors but this "
+    #        "may cause performance issues.")
+    # with must_warn(FutureWarning, msg=msg):
+    #     from_frame(df, fill_value=-1)
+    res = from_frame(df, fill_value=-1)
+    expected_data = [[['a', 1],
+                      ['a', 1]],
+                     [['a', 1],
+                      [-1, -1]]]
+    expected_axes = [
+        # third axis is anonymous
+        Axis('idx1=0,1'), Axis('idx2=a,b'), Axis('col1,col2')
+    ]
+    expected = Array(expected_data, expected_axes, dtype=object)
+    assert_larray_equal(res, expected)
+
 
 def test_asarray():
     series = pd.Series([0, 1, 2], ['a0', 'a1', 'a2'], name='a')