Fix splitter when table is only one row wide (#8839)

Author: sjrl

haystack/components/preprocessors/csv_document_splitter.py

@@ -195,7 +195,7 @@ def _split_dataframe(
         df_length = df.shape[0] if axis == "row" else df.shape[1]
         for empty_start_idx, empty_end_idx in split_indices + [(df_length, df_length)]:
             # Avoid empty splits
-            if empty_start_idx - table_start_idx > 1:
+            if empty_start_idx - table_start_idx >= 1:
                 if axis == "row":
                     sub_table = df.iloc[table_start_idx:empty_start_idx]
                 else:

test/components/preprocessors/test_csv_document_splitter.py

@@ -227,6 +227,12 @@ def test_csv_with_blank_lines(self, splitter: CSVDocumentSplitter) -> None:
             assert table.content == expected_tables[i]
             assert table.meta == expected_meta[i]
 
+    def test_sub_table_with_one_row(self):
+        splitter = CSVDocumentSplitter(row_split_threshold=1)
+        doc = Document(content="""A,B,C\n1,2,3\n,,\n4,5,6""")
+        split_result = splitter.run([doc])
+        assert len(split_result["documents"]) == 2
+
     def test_threshold_no_effect(self, two_tables_sep_by_two_empty_rows: str) -> None:
         splitter = CSVDocumentSplitter(row_split_threshold=3)
         doc = Document(content=two_tables_sep_by_two_empty_rows)