Coverage Report

Created: 2026-04-17 23:17

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
be/src/storage/segment/column_reader_cache.cpp
Line
Count
Source
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
#include "storage/segment/column_reader_cache.h"
19
20
#include "storage/segment/column_meta_accessor.h"
21
#include "storage/segment/segment.h"
22
#include "storage/segment/variant/variant_column_reader.h"
23
24
bvar::Adder<int64_t> g_segment_column_reader_cache_count("segment_column_cache_count");
25
bvar::Adder<int64_t> g_segment_column_cache_hit_count("segment_column_cache_hit_count");
26
bvar::Adder<int64_t> g_segment_column_cache_miss_count("segment_column_cache_miss_count");
27
bvar::Adder<int64_t> g_segment_column_cache_evict_count("segment_column_cache_evict_count");
28
29
namespace doris::segment_v2 {
30
31
ColumnReaderCache::ColumnReaderCache(
32
        ColumnMetaAccessor* accessor, TabletSchemaSPtr tablet_schema,
33
        io::FileReaderSPtr file_reader, uint64_t num_rows,
34
        std::function<Status(std::shared_ptr<SegmentFooterPB>&, OlapReaderStatistics*)>
35
                get_footer_cb)
36
5.73k
        : _accessor(accessor),
37
5.73k
          _tablet_schema(std::move(tablet_schema)),
38
5.73k
          _file_reader(std::move(file_reader)),
39
5.73k
          _num_rows(num_rows),
40
5.73k
          _get_footer_cb(std::move(get_footer_cb)) {}
41
42
5.73k
ColumnReaderCache::~ColumnReaderCache() {
43
5.73k
    g_segment_column_reader_cache_count << -_cache_map.size();
44
5.73k
}
45
46
17.4k
std::shared_ptr<ColumnReader> ColumnReaderCache::_lookup(const ColumnReaderCacheKey& key) {
47
17.4k
    std::lock_guard<std::mutex> lock(_cache_mutex);
48
17.4k
    auto it = _cache_map.find(key);
49
17.4k
    if (it == _cache_map.end()) {
50
13.4k
        g_segment_column_cache_miss_count << 1;
51
13.4k
        return nullptr;
52
13.4k
    }
53
    // Move the accessed node to the front of the linked list
54
3.97k
    _lru_list.splice(_lru_list.begin(), _lru_list, it->second);
55
3.97k
    DCHECK_EQ(it->second->key.first, key.first);
56
3.97k
    g_segment_column_cache_hit_count << 1;
57
3.97k
    return it->second->reader;
58
17.4k
}
59
60
void ColumnReaderCache::_insert_locked_nocheck(const ColumnReaderCacheKey& key,
61
13.4k
                                               const std::shared_ptr<ColumnReader>& reader) {
62
    // If capacity exceeded, remove least recently used (tail)
63
13.4k
    if (_cache_map.size() >= config::max_segment_partial_column_cache_size) {
64
220
        g_segment_column_reader_cache_count << -1;
65
220
        g_segment_column_cache_evict_count << 1;
66
220
        auto last_it = _lru_list.end();
67
220
        --last_it;
68
220
        _cache_map.erase(last_it->key);
69
220
        _lru_list.pop_back();
70
220
    }
71
13.4k
    g_segment_column_reader_cache_count << 1;
72
13.4k
    _lru_list.push_front(CacheNode {
73
13.4k
            .key = key, .reader = reader, .last_access = std::chrono::steady_clock::now()});
74
13.4k
    _cache_map[key] = _lru_list.begin();
75
13.4k
}
76
77
void ColumnReaderCache::_insert_direct(const ColumnReaderCacheKey& key,
78
13.4k
                                       const std::shared_ptr<ColumnReader>& column_reader) {
79
13.4k
    std::lock_guard<std::mutex> lock(_cache_mutex);
80
13.4k
    _insert_locked_nocheck(key, column_reader);
81
13.4k
}
82
83
std::map<int32_t, std::shared_ptr<ColumnReader>> ColumnReaderCache::get_available_readers(
84
7
        bool include_subcolumns) {
85
7
    std::lock_guard<std::mutex> lock(_cache_mutex);
86
7
    std::map<int32_t, std::shared_ptr<ColumnReader>> readers;
87
12
    for (const auto& node : _lru_list) {
88
12
        if (include_subcolumns || node.key.second.empty()) {
89
12
            readers.insert({node.key.first, node.reader});
90
12
        }
91
12
    }
92
7
    return readers;
93
7
}
94
95
Status ColumnReaderCache::get_column_reader(int32_t col_uid,
96
                                            std::shared_ptr<ColumnReader>* column_reader,
97
17.1k
                                            OlapReaderStatistics* stats) {
98
    // Attempt to find in cache
99
17.1k
    if (auto cached = _lookup({col_uid, {}})) {
100
3.97k
        *column_reader = cached;
101
3.97k
        return Status::OK();
102
3.97k
    }
103
    // Load footer once under cache mutex (not thread-safe otherwise)
104
13.1k
    std::shared_ptr<SegmentFooterPB> footer_pb_shared;
105
13.1k
    {
106
13.1k
        std::lock_guard<std::mutex> lock(_cache_mutex);
107
13.1k
        RETURN_IF_ERROR(_get_footer_cb(footer_pb_shared, stats));
108
13.1k
    }
109
110
    // Lookup column meta by uid via ColumnMetaAccessor. If not initialized or not found, return NOT_FOUND.
111
13.1k
    ColumnMetaPB meta;
112
13.1k
    Status st_meta = _accessor->get_column_meta_by_uid(*footer_pb_shared, col_uid, &meta);
113
13.1k
    if (st_meta.is<ErrorCode::NOT_FOUND>()) {
114
0
        *column_reader = nullptr;
115
0
        return st_meta;
116
0
    }
117
13.1k
    RETURN_IF_ERROR(st_meta);
118
119
13.1k
    ColumnReaderOptions opts {.kept_in_memory = _tablet_schema->is_in_memory(),
120
13.1k
                              .be_exec_version = _be_exec_version,
121
13.1k
                              .tablet_schema = _tablet_schema};
122
123
13.1k
    std::shared_ptr<ColumnReader> reader;
124
13.1k
    if ((FieldType)meta.type() == FieldType::OLAP_FIELD_TYPE_VARIANT) {
125
        // Variant root columns require VariantColumnReader, which encapsulates
126
        // subcolumn layout, sparse columns and external meta.
127
363
        std::unique_ptr<VariantColumnReader> variant_reader(new VariantColumnReader());
128
363
        RETURN_IF_ERROR(variant_reader->init(opts, _accessor, footer_pb_shared, col_uid, _num_rows,
129
363
                                             _file_reader));
130
363
        reader.reset(variant_reader.release());
131
363
        VLOG_DEBUG << "insert cache (variant): uid=" << col_uid << " col_id=" << meta.column_id();
132
12.7k
    } else {
133
        // For non-variant columns, we can create reader directly from ColumnMetaPB.
134
18.4E
        VLOG_DEBUG << "insert cache: uid=" << col_uid << " col_id=" << meta.column_id();
135
12.7k
        RETURN_IF_ERROR(ColumnReader::create(opts, meta, _num_rows, _file_reader, &reader));
136
12.7k
    }
137
138
13.1k
    _insert_direct({col_uid, {}}, reader);
139
13.1k
    *column_reader = std::move(reader);
140
13.1k
    return Status::OK();
141
13.1k
}
142
143
Status ColumnReaderCache::get_path_column_reader(int32_t col_uid, PathInData relative_path,
144
                                                 std::shared_ptr<ColumnReader>* column_reader,
145
                                                 OlapReaderStatistics* stats,
146
306
                                                 const SubcolumnColumnMetaInfo::Node* node_hint) {
147
    // Attempt to find in cache first
148
306
    if (auto cached = _lookup({col_uid, relative_path})) {
149
0
        *column_reader = cached;
150
0
        return Status::OK();
151
0
    }
152
153
306
    if (!_accessor->has_column_uid(col_uid)) {
154
0
        *column_reader = nullptr;
155
0
        return Status::Error<ErrorCode::NOT_FOUND, false>("column not found in segment, col_uid={}",
156
0
                                                          col_uid);
157
0
    }
158
159
    // Load footer once under cache mutex (not thread-safe otherwise)
160
306
    std::shared_ptr<SegmentFooterPB> footer_pb_shared;
161
306
    {
162
306
        std::lock_guard<std::mutex> lock(_cache_mutex);
163
306
        RETURN_IF_ERROR(_get_footer_cb(footer_pb_shared, stats));
164
306
    }
165
166
    // Ensure variant root reader is available in cache.
167
306
    ColumnReaderOptions opts {.kept_in_memory = _tablet_schema->is_in_memory(),
168
306
                              .be_exec_version = _be_exec_version,
169
306
                              .tablet_schema = _tablet_schema};
170
306
    std::shared_ptr<ColumnReader> variant_column_reader;
171
306
    RETURN_IF_ERROR(get_column_reader(col_uid, &variant_column_reader, stats));
172
173
306
    if (relative_path.empty()) {
174
0
        *column_reader = std::move(variant_column_reader);
175
0
        return Status::OK();
176
0
    }
177
178
    // Delegate path-level reader creation to VariantColumnReader, which hides
179
    // inline vs external meta details.
180
306
    std::shared_ptr<ColumnReader> path_reader;
181
306
    auto* vreader = static_cast<VariantColumnReader*>(variant_column_reader.get());
182
306
    Status st = vreader->create_path_reader(relative_path, opts, _accessor, *footer_pb_shared,
183
306
                                            _file_reader, _num_rows, &path_reader);
184
306
    if (st.is<ErrorCode::NOT_FOUND>()) {
185
5
        *column_reader = nullptr;
186
5
        return st;
187
5
    }
188
301
    RETURN_IF_ERROR(st);
189
190
    // Cache and return
191
301
    _insert_direct({col_uid, relative_path}, path_reader);
192
301
    *column_reader = std::move(path_reader);
193
301
    return Status::OK();
194
301
}
195
196
} // namespace doris::segment_v2