morningman commented on code in PR #14470:
URL: https://github.com/apache/doris/pull/14470#discussion_r1035958996
##########
be/src/vec/exec/format/table/iceberg_reader.cpp:
##########
@@ -122,6 +144,16 @@ Status IcebergTableReader::init_row_filters() {
auto files = table_desc.delete_files;
if (delete_file_type == POSITON_DELELE) {
// position delete
+ auto row_desc = RowDescriptor(_state->desc_tbl(),
+
std::vector<TupleId>({table_desc.delete_table_tuple_id}),
+ std::vector<bool>({false}));
+ if (_data_path_conjunct_ctx == nullptr) {
Review Comment:
No need to check `if (_data_path_conjunct_ctx == nullptr)`, it must be null
here.
##########
be/src/vec/exec/format/table/iceberg_reader.cpp:
##########
@@ -134,17 +166,32 @@ Status IcebergTableReader::init_row_filters() {
ParquetReader* delete_reader = new ParquetReader(
_profile, _params, delete_range,
_state->query_options().batch_size,
const_cast<cctz::time_zone*>(&_state->timezone_obj()));
- FileMetaData* metadata = nullptr;
- RETURN_IF_ERROR(delete_reader->file_metadata(&metadata));
-
- auto& delete_file_schema = metadata->schema();
+ if (_delete_file_schema == nullptr) {
+ FileMetaData* metadata = nullptr;
+ RETURN_IF_ERROR(delete_reader->file_metadata(&metadata));
+ if (metadata == nullptr) {
+ break;
+ }
+ _delete_file_schema.reset(&metadata->schema());
+ int num_of_col = _delete_file_schema->size();
+ for (auto i = 0; i < num_of_col; ++i) {
+ const FieldSchema* field =
_delete_file_schema->get_column(i);
+ _column_schemas.emplace_back(field);
+ }
+ }
+ DCHECK_EQ(_column_schemas.size(), _delete_file_schema->size());
vector<std::string> names;
- for (auto i = 0; i < delete_file_schema.size(); ++i) {
- const FieldSchema* field =
delete_file_schema.get_column(i);
- _column_schemas.emplace_back(field);
+ for (const FieldSchema* field : _column_schemas) {
Review Comment:
This `for (const FieldSchema* field : _column_schemas)` can also be done
only once.
Because all delete files have same column names.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
