JingsongLi commented on code in PR #11031:
URL: https://github.com/apache/paimon/pull/11031#discussion_r4522852705


##########
paimon-python/pypaimon/execution/python/read/reader/aggregate/aggregators.py:
##########
@@ -1240,8 +1263,19 @@ def _resolve_ts_field_index(self, row_type: RowType, 
options: CoreOptions, field
                     )
                 )
 
+        ts_type = row_type.fields[ts_field_index].type
+        ts_base = _atomic_base_name(ts_type)
+        if ts_base not in _KEYTIME_TIMESTAMP_TYPES:

Review Comment:
   **[P1] Keep existing PyPaimon aggregation tables readable**
   
   At `b000c3945f` I created a filesystem primary-key aggregation table with 
`updates MAP<STRING, ROW<value STRING, event_time DOUBLE>>`, set 
`merge_map_with_keytime` on `updates`, and committed two values for the same 
map key with timestamps 9.5 and 10.25. A normal PyPaimon read returned the 
latter value. Reopening the same persisted table on `55afd99d3` fails here with 
`ValueError` before reading the rows. Java previously rejected DOUBLE, but the 
Python entry point allowed it and produced readable tables. Could Python retain 
its existing comparison path for previously accepted ordered timestamp types, 
or handle the compatibility transition separately?



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to