Backport PR #36175: BUG: read_excel for ods files raising UnboundLocalError in certain cases (#36355)

Co-authored-by: Asish Mahapatra <[email protected]>

Author: simonjayhawkins

doc/source/whatsnew/v1.1.3.rst

@@ -18,6 +18,7 @@ Fixed regressions
 - Fixed regression in :class:`IntegerArray` unary plus and minus operations raising a ``TypeError`` (:issue:`36063`)
 - Fixed regression in :meth:`Series.__getitem__` incorrectly raising when the input was a tuple (:issue:`35534`)
 - Fixed regression in :meth:`Series.__getitem__` incorrectly raising when the input was a frozenset (:issue:`35747`)
+- Fixed regression in :meth:`read_excel` with `engine="odf"` caused ``UnboundLocalError`` in some cases where cells had nested child nodes (:issue:`36122`, and :issue:`35802`)
 -
 
 .. ---------------------------------------------------------------------------

pandas/io/excel/_odfreader.py

@@ -191,22 +191,24 @@ def _get_cell_string_value(self, cell) -> str:
         Find and decode OpenDocument text:s tags that represent
         a run length encoded sequence of space characters.
         """
-        from odf.element import Element, Text
+        from odf.element import Element
         from odf.namespaces import TEXTNS
-        from odf.text import P, S
+        from odf.text import S
 
-        text_p = P().qname
         text_s = S().qname
 
-        p = cell.childNodes[0]
-
         value = []
-        if p.qname == text_p:
-            for k, fragment in enumerate(p.childNodes):
-                if isinstance(fragment, Text):
-                    value.append(fragment.data)
-                elif isinstance(fragment, Element):
-                    if fragment.qname == text_s:
-                        spaces = int(fragment.attributes.get((TEXTNS, "c"), 1))
+
+        for fragment in cell.childNodes:
+            if isinstance(fragment, Element):
+                if fragment.qname == text_s:
+                    spaces = int(fragment.attributes.get((TEXTNS, "c"), 1))
                     value.append(" " * spaces)
+                else:
+                    # recursive impl needed in case of nested fragments
+                    # with multiple spaces
+                    # https://github.com/pandas-dev/pandas/pull/36175#discussion_r484639704
+                    value.append(self._get_cell_string_value(fragment))
+            else:
+                value.append(str(fragment))
         return "".join(value)

pandas/tests/io/data/excel/gh-35802.ods

@@ -519,6 +519,23 @@ def test_reader_spaces(self, read_ext):
         )
         tm.assert_frame_equal(actual, expected)
 
+    # gh-36122, gh-35802
+    @pytest.mark.parametrize(
+        "basename,expected",
+        [
+            ("gh-35802", DataFrame({"COLUMN": ["Test (1)"]})),
+            ("gh-36122", DataFrame(columns=["got 2nd sa"])),
+        ],
+    )
+    def test_read_excel_ods_nested_xml(self, read_ext, basename, expected):
+        # see gh-35802
+        engine = pd.read_excel.keywords["engine"]
+        if engine != "odf":
+            pytest.skip(f"Skipped for engine: {engine}")
+
+        actual = pd.read_excel(basename + read_ext)
+        tm.assert_frame_equal(actual, expected)
+
     def test_reading_all_sheets(self, read_ext):
         # Test reading all sheet names by setting sheet_name to None,
         # Ensure a dict is returned.