From 4752f9fa8cf82134d553e0fc9ce69fafd17b216c Mon Sep 17 00:00:00 2001
From: Andrew Chen <48723787+chuenchen309@users.noreply.github.com>
Date: Thu, 16 Jul 2026 19:14:50 +0800
Subject: [PATCH] fix(rag): ExcelKnowledge._load must handle headerless sheets
 and multi-sheet workbooks (#3137)

Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
---
 .../src/dbgpt_ext/rag/knowledge/excel.py      |  3 +-
 .../rag/knowledge/tests/test_excel.py         | 44 +++++++++++++++++++
 2 files changed, 46 insertions(+), 1 deletion(-)
 create mode 100644 packages/dbgpt-ext/src/dbgpt_ext/rag/knowledge/tests/test_excel.py

diff --git a/packages/dbgpt-ext/src/dbgpt_ext/rag/knowledge/excel.py b/packages/dbgpt-ext/src/dbgpt_ext/rag/knowledge/excel.py
index d71167a1d..3dc985577 100644
--- a/packages/dbgpt-ext/src/dbgpt_ext/rag/knowledge/excel.py
+++ b/packages/dbgpt-ext/src/dbgpt_ext/rag/knowledge/excel.py
@@ -175,6 +175,7 @@ class ExcelKnowledge(Knowledge):
                 )
             else:
                 data_rows = unmerged_data
+                df_start_row_index = 0
                 logging.info(
                     f"Sheet '{sheet_name}': No clear header row found. "
                     f"All rows considered data."
@@ -261,7 +262,7 @@ class ExcelKnowledge(Knowledge):
                 )
                 docs.append(doc)
 
-            return docs
+        return docs
 
     @classmethod
     def support_chunk_strategy(cls) -> List[ChunkStrategy]:
diff --git a/packages/dbgpt-ext/src/dbgpt_ext/rag/knowledge/tests/test_excel.py b/packages/dbgpt-ext/src/dbgpt_ext/rag/knowledge/tests/test_excel.py
new file mode 100644
index 000000000..5370ae48e
--- /dev/null
+++ b/packages/dbgpt-ext/src/dbgpt_ext/rag/knowledge/tests/test_excel.py
@@ -0,0 +1,44 @@
+import openpyxl
+import pytest
+
+from ..excel import ExcelKnowledge
+
+
+@pytest.fixture
+def multi_sheet_xlsx(tmp_path):
+    path = tmp_path / "multi_sheet.xlsx"
+    wb = openpyxl.Workbook()
+    ws1 = wb.active
+    ws1.title = "Sheet1"
+    ws1.append(["name", "value"])
+    ws1.append(["a", 1])
+    ws2 = wb.create_sheet("Sheet2")
+    ws2.append(["name", "value"])
+    ws2.append(["b", 2])
+    wb.save(path)
+    return str(path)
+
+
+@pytest.fixture
+def numeric_only_xlsx(tmp_path):
+    path = tmp_path / "numeric_only.xlsx"
+    wb = openpyxl.Workbook()
+    ws = wb.active
+    ws.title = "NumericSheet"
+    for r in range(6):
+        ws.append([r, r + 1, r + 2])
+    wb.save(path)
+    return str(path)
+
+
+def test_load_reads_all_sheets(multi_sheet_xlsx):
+    knowledge = ExcelKnowledge(file_path=multi_sheet_xlsx)
+    docs = knowledge._load()
+    sheets_seen = {doc.metadata["sheet_name"] for doc in docs}
+    assert sheets_seen == {"Sheet1", "Sheet2"}
+
+
+def test_load_does_not_crash_without_header_row(numeric_only_xlsx):
+    knowledge = ExcelKnowledge(file_path=numeric_only_xlsx)
+    docs = knowledge._load()
+    assert len(docs) == 6