/root/doris/be/src/util/key_util.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 <gen_cpp/segment_v2.pb.h> |
21 | | |
22 | | #include <cstdint> |
23 | | #include <iterator> |
24 | | #include <string> |
25 | | #include <vector> |
26 | | |
27 | | #include "common/status.h" |
28 | | #include "util/debug_util.h" |
29 | | #include "util/faststring.h" |
30 | | #include "util/slice.h" |
31 | | |
32 | | namespace doris { |
33 | | |
34 | | // In our system, we have more complicated situation. |
35 | | // First, our keys can be nullptr. |
36 | | // Second, when key columns are not complete we want to distinguish GT and GE. For example, |
37 | | // there are two key columns a and b, we have only one condition a > 1. We can only encode |
38 | | // a prefix key 1, which is less than 1|2. This will make our read more data than |
39 | | // we actually need. So we want to add more marker. |
40 | | // a > 1: will be encoded into 1|\xFF |
41 | | // a >= 1: will be encoded into 1|\x00 |
42 | | // a = 1 and b > 1: will be encoded into 1|\x02|1 |
43 | | // a = 1 and b is null: will be encoded into 1|\x01 |
44 | | |
45 | | // Used to represent minimal value for that field |
46 | | constexpr uint8_t KEY_MINIMAL_MARKER = 0x00; |
47 | | // Used to represent a null field, which value is seemed as minimal than other values |
48 | | constexpr uint8_t KEY_NULL_FIRST_MARKER = 0x01; |
49 | | // Used to represent a normal field, which content is encoded after this marker |
50 | | constexpr uint8_t KEY_NORMAL_MARKER = 0x02; |
51 | | // Used to represent maximal value for that field |
52 | | constexpr uint8_t KEY_MAXIMAL_MARKER = 0xFF; |
53 | | // Used to represent a value greater than the normal marker by 1, using by MoW |
54 | | constexpr uint8_t KEY_NORMAL_NEXT_MARKER = 0x03; |
55 | | |
56 | | // Encode one row into binary according given num_keys. |
57 | | // A cell will be encoded in the format of a marker and encoded content. |
58 | | // When function encoding row, if any cell isn't found in row, this function will |
59 | | // fill a marker and return. If padding_minimal is true, KEY_MINIMAL_MARKER will |
60 | | // be added, if padding_minimal is false, KEY_MAXIMAL_MARKER will be added. |
61 | | // If all num_keys are found in row, no marker will be added. |
62 | | template <typename RowType, bool is_mow = false> |
63 | | void encode_key_with_padding(std::string* buf, const RowType& row, size_t num_keys, |
64 | 2 | bool padding_minimal) { |
65 | 6 | for (auto cid = 0; cid < num_keys; cid++) { |
66 | 6 | auto field = row.schema()->column(cid); |
67 | 6 | if (field == nullptr) { |
68 | 2 | if (padding_minimal) { |
69 | 1 | buf->push_back(KEY_MINIMAL_MARKER); |
70 | 1 | } else { |
71 | 1 | if (is_mow) { |
72 | 0 | buf->push_back(KEY_NORMAL_NEXT_MARKER); |
73 | 1 | } else { |
74 | 1 | buf->push_back(KEY_MAXIMAL_MARKER); |
75 | 1 | } |
76 | 1 | } |
77 | 2 | break; |
78 | 2 | } |
79 | | |
80 | 4 | auto cell = row.cell(cid); |
81 | 4 | if (cell.is_null()) { |
82 | 1 | buf->push_back(KEY_NULL_FIRST_MARKER); |
83 | 1 | continue; |
84 | 1 | } |
85 | 3 | buf->push_back(KEY_NORMAL_MARKER); |
86 | 3 | if (is_mow) { |
87 | 0 | field->full_encode_ascending(cell.cell_ptr(), buf); |
88 | 3 | } else { |
89 | 3 | field->encode_ascending(cell.cell_ptr(), buf); |
90 | 3 | } |
91 | 3 | } |
92 | 2 | } _ZN5doris23encode_key_with_paddingINS_9RowCursorELb0EEEvPNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERKT_mb Line | Count | Source | 64 | 2 | bool padding_minimal) { | 65 | 6 | for (auto cid = 0; cid < num_keys; cid++) { | 66 | 6 | auto field = row.schema()->column(cid); | 67 | 6 | if (field == nullptr) { | 68 | 2 | if (padding_minimal) { | 69 | 1 | buf->push_back(KEY_MINIMAL_MARKER); | 70 | 1 | } else { | 71 | 1 | if (is_mow) { | 72 | 0 | buf->push_back(KEY_NORMAL_NEXT_MARKER); | 73 | 1 | } else { | 74 | 1 | buf->push_back(KEY_MAXIMAL_MARKER); | 75 | 1 | } | 76 | 1 | } | 77 | 2 | break; | 78 | 2 | } | 79 | | | 80 | 4 | auto cell = row.cell(cid); | 81 | 4 | if (cell.is_null()) { | 82 | 1 | buf->push_back(KEY_NULL_FIRST_MARKER); | 83 | 1 | continue; | 84 | 1 | } | 85 | 3 | buf->push_back(KEY_NORMAL_MARKER); | 86 | 3 | if (is_mow) { | 87 | 0 | field->full_encode_ascending(cell.cell_ptr(), buf); | 88 | 3 | } else { | 89 | 3 | field->encode_ascending(cell.cell_ptr(), buf); | 90 | 3 | } | 91 | 3 | } | 92 | 2 | } |
Unexecuted instantiation: _ZN5doris23encode_key_with_paddingINS_9RowCursorELb1EEEvPNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERKT_mb |
93 | | |
94 | | // Encode one row into binary according given num_keys. |
95 | | // Client call this function must assure that row contains the first |
96 | | // num_keys columns. |
97 | | template <typename RowType, bool full_encode = false> |
98 | 128k | void encode_key(std::string* buf, const RowType& row, size_t num_keys) { |
99 | 386k | for (auto cid = 0; cid < num_keys; cid++) { |
100 | 257k | auto cell = row.cell(cid); |
101 | 257k | if (cell.is_null()) { |
102 | 1 | buf->push_back(KEY_NULL_FIRST_MARKER); |
103 | 1 | continue; |
104 | 1 | } |
105 | 257k | buf->push_back(KEY_NORMAL_MARKER); |
106 | 257k | if (full_encode) { |
107 | 257k | row.schema()->column(cid)->full_encode_ascending(cell.cell_ptr(), buf); |
108 | 257k | } else { |
109 | 1 | row.schema()->column(cid)->encode_ascending(cell.cell_ptr(), buf); |
110 | 1 | } |
111 | 257k | } |
112 | 128k | } _ZN5doris10encode_keyINS_9RowCursorELb0EEEvPNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERKT_m Line | Count | Source | 98 | 1 | void encode_key(std::string* buf, const RowType& row, size_t num_keys) { | 99 | 3 | for (auto cid = 0; cid < num_keys; cid++) { | 100 | 2 | auto cell = row.cell(cid); | 101 | 2 | if (cell.is_null()) { | 102 | 1 | buf->push_back(KEY_NULL_FIRST_MARKER); | 103 | 1 | continue; | 104 | 1 | } | 105 | 1 | buf->push_back(KEY_NORMAL_MARKER); | 106 | 1 | if (full_encode) { | 107 | 0 | row.schema()->column(cid)->full_encode_ascending(cell.cell_ptr(), buf); | 108 | 1 | } else { | 109 | 1 | row.schema()->column(cid)->encode_ascending(cell.cell_ptr(), buf); | 110 | 1 | } | 111 | 1 | } | 112 | 1 | } |
_ZN5doris10encode_keyINS_9RowCursorELb1EEEvPNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEERKT_m Line | Count | Source | 98 | 128k | void encode_key(std::string* buf, const RowType& row, size_t num_keys) { | 99 | 386k | for (auto cid = 0; cid < num_keys; cid++) { | 100 | 257k | auto cell = row.cell(cid); | 101 | 257k | if (cell.is_null()) { | 102 | 0 | buf->push_back(KEY_NULL_FIRST_MARKER); | 103 | 0 | continue; | 104 | 0 | } | 105 | 257k | buf->push_back(KEY_NORMAL_MARKER); | 106 | 257k | if (full_encode) { | 107 | 257k | row.schema()->column(cid)->full_encode_ascending(cell.cell_ptr(), buf); | 108 | 257k | } else { | 109 | 0 | row.schema()->column(cid)->encode_ascending(cell.cell_ptr(), buf); | 110 | 0 | } | 111 | 257k | } | 112 | 128k | } |
|
113 | | |
114 | | } // namespace doris |