fix: keep an XLSX cell from breaking the markdown table

haystack/components/converters/xlsx.py

@@ -234,6 +234,15 @@ def _extract_tables(self, bytestream: ByteStream) -> tuple[list[str], list[dict]
                     "missingval": "",
                     **self.table_format_kwargs,
                 }
+                if resolved_kwargs["tablefmt"] == "pipe":
+                    value = value.replace(
+                        {
+                            r"\r\n|\r|\n": " ",  # keep in-cell line breaks from creating extra Markdown rows
+                            r"(\\*)\|": r"\1\1\\|",  # escape pipes but preserve any preceding literal backslashes
+                        },
+                        regex=True,
+                    )
+
                 # to_markdown uses tabulate, whose missingval only covers None: a NaN
                 # reaches the formatter as a number and is written out as "nan". Replace
                 # the empty cells with None so an empty cell reads as empty, the way

releasenotes/notes/xlsx-markdown-escape-25b936b365fd48f1.yaml

@@ -0,0 +1,6 @@
+---
+fixes:
+  - |
+    ``XLSXToDocument`` now escapes pipe characters and replaces in-cell line breaks
+    with spaces in the default Markdown pipe table output. This keeps cell content
+    from being interpreted as additional table columns or rows.

test/components/converters/test_xlsx_to_document.py

@@ -108,6 +108,31 @@ def test_run_markdown_missing_value(self, test_files_path: Path) -> None:
             == "|    | A     | B     |\n|---:|:------|:------|\n|  1 | col_c | col_d |\n|  2 | True  | N/A   |"
         )
 
+    @pytest.mark.parametrize(
+        ("cell_value", "expected_row"),
+        [
+            pytest.param("a|b", "|  1 | a\\|b |", id="pipe"),
+            pytest.param(r"a\|b", r"|  1 | a\\\|b |", id="backslash-before-pipe"),
+            pytest.param("first\r\nsecond", "|  1 | first second |", id="windows-line-break"),
+            pytest.param("first line\nsecond line", "|  1 | first line second line |", id="line-break"),
+        ],
+    )
+    def test_run_markdown_escapes_cell_content(self, tmp_path: Path, cell_value: str, expected_row: str) -> None:
+        """A pipe would be read as a column separator, and a line break would end the row in the middle."""
+        workbook = Workbook()
+        sheet = workbook.active
+        assert sheet is not None
+        sheet["A1"] = cell_value
+        path = tmp_path / "cell.xlsx"
+        workbook.save(path)
+
+        content = XLSXToDocument(table_format="markdown").run(sources=[path])["documents"][0].content
+        assert content is not None
+        rows = content.split("\n")
+
+        assert len(rows) == 3
+        assert rows[2] == expected_row
+
     @pytest.mark.parametrize(
         "sheet_name, expected_sheet_name, expected_content",
         [