yiguolei commented on code in PR #35432:
URL: https://github.com/apache/doris/pull/35432#discussion_r1615665538
##########
be/src/olap/rowset/beta_rowset_reader.cpp:
##########
@@ -249,38 +250,60 @@ Status
BetaRowsetReader::get_segment_iterators(RowsetReaderContext* read_context
// load segments
bool should_use_cache = use_cache || _read_context->reader_type ==
ReaderType::READER_QUERY;
- RETURN_IF_ERROR(SegmentLoader::instance()->load_segments(_rowset,
&_segment_cache_handle,
+ SegmentCacheHandle segment_cache_handle;
+ RETURN_IF_ERROR(SegmentLoader::instance()->load_segments(_rowset,
&segment_cache_handle,
should_use_cache));
// create iterator for each segment
- auto& segments = _segment_cache_handle.get_segments();
+ auto& segments = segment_cache_handle.get_segments();
+ _segments_rows.resize(segments.size());
+ for (size_t i = 0; i < segments.size(); i++) {
+ _segments_rows[i] = segments[i]->num_rows();
+ }
+
auto [seg_start, seg_end] = _segment_offsets;
if (seg_start == seg_end) {
seg_start = 0;
seg_end = segments.size();
}
+ const bool is_merge_iterator = _is_merge_iterator();
for (int i = seg_start; i < seg_end; i++) {
auto& seg_ptr = segments[i];
std::unique_ptr<RowwiseIterator> iter;
- Status status;
- /// If `_segment_row_ranges` is empty, the segment is not split.
- if (_segment_row_ranges.empty()) {
- _read_options.row_ranges.clear();
- status = seg_ptr->new_iterator(_input_schema, _read_options,
&iter);
+ if (is_merge_iterator) {
+ Status status;
+ /// If `_segment_row_ranges` is empty, the segment is not split.
+ if (_segment_row_ranges.empty()) {
+ _read_options.row_ranges.clear();
+ status = seg_ptr->new_iterator(_input_schema, _read_options,
&iter);
+ } else {
+ DCHECK_EQ(seg_end - seg_start, _segment_row_ranges.size());
+ auto local_options = _read_options;
+ local_options.row_ranges = _segment_row_ranges[i - seg_start];
+ status = seg_ptr->new_iterator(_input_schema, local_options,
&iter);
+ }
+
+ if (!status.ok()) {
+ LOG(WARNING) << "failed to create iterator[" << seg_ptr->id()
+ << "]: " << status.to_string();
+ return Status::Error<ROWSET_READER_INIT>(status.to_string());
+ }
} else {
- DCHECK_EQ(seg_end - seg_start, _segment_row_ranges.size());
- auto local_options = _read_options;
- local_options.row_ranges = _segment_row_ranges[i - seg_start];
- status = seg_ptr->new_iterator(_input_schema, local_options,
&iter);
+ if (_segment_row_ranges.empty()) {
+ _read_options.row_ranges.clear();
+ iter = std::make_unique<LazyInitSegmentIterator>(seg_ptr,
_input_schema,
Review Comment:
Add some comment here to explain why we need lazy init segment iterator.
--
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: commits-unsubscr...@doris.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org
For additional commands, e-mail: commits-h...@doris.apache.org
