/root/doris/be/src/olap/rowid_conversion.h
Line | Count | Source (jump to first uncovered line) |
1 | | // Licensed to the Apache Software Foundation (ASF) under one |
2 | | // or more contributor license agreements. See the NOTICE file |
3 | | // distributed with this work for additional information |
4 | | // regarding copyright ownership. The ASF licenses this file |
5 | | // to you under the Apache License, Version 2.0 (the |
6 | | // "License"); you may not use this file except in compliance |
7 | | // with the License. You may obtain a copy of the License at |
8 | | // |
9 | | // http://www.apache.org/licenses/LICENSE-2.0 |
10 | | // |
11 | | // Unless required by applicable law or agreed to in writing, |
12 | | // software distributed under the License is distributed on an |
13 | | // "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY |
14 | | // KIND, either express or implied. See the License for the |
15 | | // specific language governing permissions and limitations |
16 | | // under the License. |
17 | | |
18 | | #pragma once |
19 | | |
20 | | #include <map> |
21 | | #include <vector> |
22 | | |
23 | | #include "olap/olap_common.h" |
24 | | #include "olap/utils.h" |
25 | | #include "runtime/thread_context.h" |
26 | | |
27 | | namespace doris { |
28 | | |
29 | | // For unique key merge on write table, we should update delete bitmap |
30 | | // of destination rowset when compaction finished. |
31 | | // Through the row id correspondence between the source rowset and the |
32 | | // destination rowset, we can quickly update the delete bitmap of the |
33 | | // destination rowset. |
34 | | class RowIdConversion { |
35 | | public: |
36 | 113 | RowIdConversion() = default; |
37 | 113 | ~RowIdConversion() { RELEASE_THREAD_MEM_TRACKER(_seg_rowid_map_mem_used); } |
38 | | |
39 | | // resize segment rowid map to its rows num |
40 | 292 | void init_segment_map(const RowsetId& src_rowset_id, const std::vector<uint32_t>& num_rows) { |
41 | 292 | size_t delta_std_pair_cap = 0; |
42 | 734 | for (size_t i = 0; i < num_rows.size(); i++) { |
43 | 442 | uint32_t id = _segments_rowid_map.size(); |
44 | 442 | _segment_to_id_map.emplace(std::pair<RowsetId, uint32_t> {src_rowset_id, i}, id); |
45 | 442 | _id_to_segment_map.emplace_back(src_rowset_id, i); |
46 | 442 | std::vector<std::pair<uint32_t, uint32_t>> vec( |
47 | 442 | num_rows[i], std::pair<uint32_t, uint32_t>(UINT32_MAX, UINT32_MAX)); |
48 | 442 | delta_std_pair_cap += vec.capacity(); |
49 | 442 | _segments_rowid_map.emplace_back(std::move(vec)); |
50 | 442 | } |
51 | | //NOTE: manually count _segments_rowid_map's memory here, because _segments_rowid_map could be used by indexCompaction. |
52 | | // indexCompaction is a thridparty code, it's too complex to modify it. |
53 | | // refer compact_column. |
54 | 292 | track_mem_usage(delta_std_pair_cap); |
55 | 292 | } |
56 | | |
57 | | // set dst rowset id |
58 | 98 | void set_dst_rowset_id(const RowsetId& dst_rowset_id) { _dst_rowst_id = dst_rowset_id; } |
59 | 2 | const RowsetId get_dst_rowset_id() { return _dst_rowst_id; } |
60 | | |
61 | | // add row id to the map |
62 | | void add(const std::vector<RowLocation>& rss_row_ids, |
63 | 5.15k | const std::vector<uint32_t>& dst_segments_num_row) { |
64 | 4.68M | for (auto& item : rss_row_ids) { |
65 | 4.68M | if (item.row_id == -1) { |
66 | 0 | continue; |
67 | 0 | } |
68 | 4.68M | uint32_t id = _segment_to_id_map.at( |
69 | 4.68M | std::pair<RowsetId, uint32_t> {item.rowset_id, item.segment_id}); |
70 | 4.68M | if (_cur_dst_segment_id < dst_segments_num_row.size() && |
71 | 4.68M | _cur_dst_segment_rowid >= dst_segments_num_row[_cur_dst_segment_id]) { |
72 | 3.63k | _cur_dst_segment_id++; |
73 | 3.63k | _cur_dst_segment_rowid = 0; |
74 | 3.63k | } |
75 | 4.68M | _segments_rowid_map[id][item.row_id] = |
76 | 4.68M | std::pair<uint32_t, uint32_t> {_cur_dst_segment_id, _cur_dst_segment_rowid++}; |
77 | 4.68M | } |
78 | 5.15k | } |
79 | | |
80 | | // get destination RowLocation |
81 | | // return non-zero if the src RowLocation does not exist |
82 | 1.75M | int get(const RowLocation& src, RowLocation* dst) const { |
83 | 1.75M | auto iter = _segment_to_id_map.find({src.rowset_id, src.segment_id}); |
84 | 1.75M | if (iter == _segment_to_id_map.end()) { |
85 | 1 | return -1; |
86 | 1 | } |
87 | 1.75M | const auto& rowid_map = _segments_rowid_map[iter->second]; |
88 | 1.75M | if (src.row_id >= rowid_map.size()) { |
89 | 1 | return -1; |
90 | 1 | } |
91 | 1.75M | auto& [dst_segment_id, dst_rowid] = rowid_map[src.row_id]; |
92 | 1.75M | if (dst_segment_id == UINT32_MAX && dst_rowid == UINT32_MAX) { |
93 | 764k | return -1; |
94 | 764k | } |
95 | | |
96 | 993k | dst->rowset_id = _dst_rowst_id; |
97 | 993k | dst->segment_id = dst_segment_id; |
98 | 993k | dst->row_id = dst_rowid; |
99 | 993k | return 0; |
100 | 1.75M | } |
101 | | |
102 | | const std::vector<std::vector<std::pair<uint32_t, uint32_t>>>& get_rowid_conversion_map() |
103 | 2 | const { |
104 | 2 | return _segments_rowid_map; |
105 | 2 | } |
106 | | |
107 | 2 | const std::map<std::pair<RowsetId, uint32_t>, uint32_t>& get_src_segment_to_id_map() { |
108 | 2 | return _segment_to_id_map; |
109 | 2 | } |
110 | | |
111 | 0 | std::pair<RowsetId, uint32_t> get_segment_by_id(uint32_t id) const { |
112 | 0 | DCHECK_GT(_id_to_segment_map.size(), id); |
113 | 0 | return _id_to_segment_map.at(id); |
114 | 0 | } |
115 | | |
116 | 0 | uint32_t get_id_by_segment(const std::pair<RowsetId, uint32_t>& segment) const { |
117 | 0 | return _segment_to_id_map.at(segment); |
118 | 0 | } |
119 | | |
120 | | private: |
121 | 292 | void track_mem_usage(size_t delta_std_pair_cap) { |
122 | 292 | _std_pair_cap += delta_std_pair_cap; |
123 | | |
124 | 292 | size_t new_size = |
125 | 292 | _std_pair_cap * sizeof(std::pair<uint32_t, uint32_t>) + |
126 | 292 | _segments_rowid_map.capacity() * sizeof(std::vector<std::pair<uint32_t, uint32_t>>); |
127 | | |
128 | 292 | RELEASE_THREAD_MEM_TRACKER(_seg_rowid_map_mem_used); |
129 | 292 | CONSUME_THREAD_MEM_TRACKER(new_size); |
130 | 292 | _seg_rowid_map_mem_used = new_size; |
131 | 292 | } |
132 | | |
133 | | private: |
134 | | // the first level vector: index indicates src segment. |
135 | | // the second level vector: index indicates row id of source segment, |
136 | | // value indicates row id of destination segment. |
137 | | // <UINT32_MAX, UINT32_MAX> indicates current row not exist. |
138 | | std::vector<std::vector<std::pair<uint32_t, uint32_t>>> _segments_rowid_map; |
139 | | size_t _seg_rowid_map_mem_used {0}; |
140 | | size_t _std_pair_cap {0}; |
141 | | |
142 | | // Map source segment to 0 to n |
143 | | std::map<std::pair<RowsetId, uint32_t>, uint32_t> _segment_to_id_map; |
144 | | |
145 | | // Map 0 to n to source segment |
146 | | std::vector<std::pair<RowsetId, uint32_t>> _id_to_segment_map; |
147 | | |
148 | | // dst rowset id |
149 | | RowsetId _dst_rowst_id; |
150 | | |
151 | | // current dst segment id |
152 | | std::uint32_t _cur_dst_segment_id = 0; |
153 | | |
154 | | // current rowid of dst segment |
155 | | std::uint32_t _cur_dst_segment_rowid = 0; |
156 | | }; |
157 | | |
158 | | } // namespace doris |