This is an automated email from the ASF dual-hosted git repository.
morningman pushed a commit to branch branch-1.2-lts
in repository https://gitbox.apache.org/repos/asf/doris.git
commit 95aab248b03ed82b384cf0c2d3510f07b944dad3
Author: xueweizhang <zxw520bl...@163.com>
AuthorDate: Mon Dec 19 10:08:38 2022 +0800
[fix](merge-on-write) delete all rows with same key in all pre segments
(#14995)
---
be/src/olap/tablet.cpp | 51 ++++++++++++++++++++++++++------------------------
be/src/olap/tablet.h | 3 ++-
2 files changed, 29 insertions(+), 25 deletions(-)
diff --git a/be/src/olap/tablet.cpp b/be/src/olap/tablet.cpp
index 11d125035e..e15bd61610 100644
--- a/be/src/olap/tablet.cpp
+++ b/be/src/olap/tablet.cpp
@@ -2032,10 +2032,10 @@ Status Tablet::calc_delete_bitmap(RowsetId rowset_id,
RowLocation loc;
// first check if exist in pre segment
if (check_pre_segments) {
- auto st = _check_pk_in_pre_segments(pre_segments, *key,
dummy_version,
- delete_bitmap, &loc);
+ auto st = _check_pk_in_pre_segments(rowset_id,
pre_segments, *key,
+ dummy_version,
delete_bitmap, &loc);
if (st.ok()) {
- delete_bitmap->add({loc.rowset_id, loc.segment_id,
dummy_version.first},
+ delete_bitmap->add({rowset_id, loc.segment_id,
dummy_version.first},
loc.row_id);
++row_id;
continue;
@@ -2045,23 +2045,27 @@ Status Tablet::calc_delete_bitmap(RowsetId rowset_id,
continue;
}
}
- auto st = lookup_row_key(*key, specified_rowset_ids, &loc,
dummy_version.first - 1);
- CHECK(st.ok() || st.is_not_found() || st.is_already_exist());
- if (st.is_not_found()) {
- ++row_id;
- continue;
- }
- // sequence id smaller than the previous one, so delete
current row
- if (st.is_already_exist()) {
- loc.rowset_id = rowset_id;
- loc.segment_id = seg->id();
- loc.row_id = row_id;
- }
+ if (!specified_rowset_ids->empty()) {
+ auto st = lookup_row_key(*key, specified_rowset_ids, &loc,
+ dummy_version.first - 1);
+ CHECK(st.ok() || st.is_not_found() ||
st.is_already_exist());
+ if (st.is_not_found()) {
+ ++row_id;
+ continue;
+ }
+ // sequence id smaller than the previous one, so delete
current row
+ if (st.is_already_exist()) {
+ loc.rowset_id = rowset_id;
+ loc.segment_id = seg->id();
+ loc.row_id = row_id;
+ }
+
+ delete_bitmap->add({loc.rowset_id, loc.segment_id,
dummy_version.first},
+ loc.row_id);
+ }
++row_id;
- delete_bitmap->add({loc.rowset_id, loc.segment_id,
dummy_version.first},
- loc.row_id);
}
remaining -= num_read;
}
@@ -2077,15 +2081,15 @@ Status Tablet::calc_delete_bitmap(RowsetId rowset_id,
}
Status Tablet::_check_pk_in_pre_segments(
- const std::vector<segment_v2::SegmentSharedPtr>& pre_segments, const
Slice& key,
- const Version& version, DeleteBitmapPtr delete_bitmap, RowLocation*
loc) {
+ RowsetId rowset_id, const std::vector<segment_v2::SegmentSharedPtr>&
pre_segments,
+ const Slice& key, const Version& version, DeleteBitmapPtr
delete_bitmap, RowLocation* loc) {
for (auto it = pre_segments.rbegin(); it != pre_segments.rend(); ++it) {
auto st = (*it)->lookup_row_key(key, loc);
CHECK(st.ok() || st.is_not_found() || st.is_already_exist());
if (st.is_not_found()) {
continue;
} else if (st.ok() && _schema->has_sequence_col() &&
- delete_bitmap->contains({loc->rowset_id, loc->segment_id,
version.first},
+ delete_bitmap->contains({rowset_id, loc->segment_id,
version.first},
loc->row_id)) {
// if has sequence col, we continue to compare the sequence_id of
// all segments, util we find an existing key.
@@ -2159,10 +2163,9 @@ Status Tablet::update_delete_bitmap(const
RowsetSharedPtr& rowset, DeleteBitmapP
for (const auto& to_del : rowset_ids_to_del) {
delete_bitmap->remove({to_del, 0, 0}, {to_del, UINT32_MAX, INT64_MAX});
}
- if (!rowset_ids_to_add.empty()) {
- RETURN_IF_ERROR(calc_delete_bitmap(rowset->rowset_id(), segments,
&rowset_ids_to_add,
- delete_bitmap, cur_version - 1,
true));
- }
+
+ RETURN_IF_ERROR(calc_delete_bitmap(rowset->rowset_id(), segments,
&rowset_ids_to_add,
+ delete_bitmap, cur_version - 1, true));
// update version without write lock, compaction and publish_txn
// will update delete bitmap, handle compaction with _rowset_update_lock
diff --git a/be/src/olap/tablet.h b/be/src/olap/tablet.h
index 318fdc8748..31e31d2958 100644
--- a/be/src/olap/tablet.h
+++ b/be/src/olap/tablet.h
@@ -385,7 +385,8 @@ private:
bool _reconstruct_version_tracker_if_necessary();
void _init_context_common_fields(RowsetWriterContext& context);
- Status _check_pk_in_pre_segments(const
std::vector<segment_v2::SegmentSharedPtr>& pre_segments,
+ Status _check_pk_in_pre_segments(RowsetId rowset_id,
+ const
std::vector<segment_v2::SegmentSharedPtr>& pre_segments,
const Slice& key, const Version& version,
DeleteBitmapPtr delete_bitmap,
RowLocation* loc);
void _rowset_ids_difference(const RowsetIdUnorderedSet& cur, const
RowsetIdUnorderedSet& pre,
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org
For additional commands, e-mail: commits-h...@doris.apache.org
