kevinjqliu commented on code in PR #1878:
URL: https://github.com/apache/iceberg-python/pull/1878#discussion_r2052855516
##########
pyiceberg/table/upsert_util.py:
##########
@@ -82,14 +82,54 @@ def get_rows_to_update(source_table: pa.Table,
target_table: pa.Table, join_cols
],
)
- return (
- source_table
- # We already know that the schema is compatible, this is to fix large_
types
- .cast(target_table.schema)
- .join(target_table, keys=list(join_cols_set), join_type="inner",
left_suffix="-lhs", right_suffix="-rhs")
- .filter(diff_expr)
- .drop_columns([f"{col}-rhs" for col in non_key_cols])
- .rename_columns({f"{col}-lhs" if col not in join_cols else col: col
for col in source_table.column_names})
- # Finally cast to the original schema since it doesn't carry
nullability:
- # https://github.com/apache/arrow/issues/45557
- ).cast(target_table.schema)
+ try:
+ return (
+ source_table
+ # We already know that the schema is compatible, this is to fix
large_ types
+ .cast(target_table.schema)
+ .join(target_table, keys=list(join_cols_set), join_type="inner",
left_suffix="-lhs", right_suffix="-rhs")
+ .filter(diff_expr)
+ .drop_columns([f"{col}-rhs" for col in non_key_cols])
+ .rename_columns({f"{col}-lhs" if col not in join_cols else col:
col for col in source_table.column_names})
+ # Finally cast to the original schema since it doesn't carry
nullability:
+ # https://github.com/apache/arrow/issues/45557
+ ).cast(target_table.schema)
+ except pa.ArrowInvalid:
+ # When we are not able to compare (e.g. due to unsupported types),
+ # fall back to selecting only rows in the source table that do NOT
already exist in the target.
+ # See: https://github.com/apache/arrow/issues/35785
+ MARKER_COLUMN_NAME = "__from_target"
+ INDEX_COLUMN_NAME = "__source_index"
+
+ if MARKER_COLUMN_NAME in join_cols_set or INDEX_COLUMN_NAME in
join_cols_set:
+ raise ValueError(
+ f"{MARKER_COLUMN_NAME} and {INDEX_COLUMN_NAME} are reserved
for joining "
+ f"DataFrames, and cannot be used as column names"
+ ) from None
+
Review Comment:
Since we cannot compare complex types, i would assume we cannot use any
complex type columns as the join keys. If thats the case, we should check if
its used and then fail
--
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: issues-unsubscr...@iceberg.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@iceberg.apache.org
For additional commands, e-mail: issues-h...@iceberg.apache.org
