Skip to content

RAGFlow is an agent-native database: a Retrieval-Augmented Generation (RAG) engine that grew into context infrastructure.

Fix: table column auto parse missing metadata (#19439)

Wang Qie4c706b536fc
diff --git a/rag/app/table.py b/rag/app/table.py
index b49f882de6..5b72eff59c 100644
--- a/rag/app/table.py
+++ b/rag/app/table.py
@@ -635,14 +635,10 @@ def chunk(filename, binary=None, from_page=0, to_page=MAXIMUM_TASK_PAGE_NUMBER,
fld = clmns_map[j][0]
if clmn_tys[j] != "text":
val = row[col_name]
- # If a string value ended up in a non-text column,
- # it was preserved from a failed conversion in
- # column_data_type. Skip storing in the typed ES
- # field to avoid type mapping errors; the value is
- # already in text_fields for chunk content.
- if isinstance(val, str):
- pass
- else:
+ # Valid datetime and bool conversions are strings;
+ # other strings in typed columns are failed conversions.
+ valid_typed_string = isinstance(val, str) and (clmn_tys[j] == "datetime" and trans_datatime(val) or clmn_tys[j] == "bool" and trans_bool(val))
+ if not isinstance(val, str) or valid_typed_string:
stored[fld] = val
else:
cell = row[col_name]