Coverage Report

Created: 2026-07-08 16:24

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
be/src/format_v2/parquet/parquet_profile.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 "format_v2/parquet/parquet_profile.h"
19
20
#include "format_v2/parquet/parquet_statistics.h"
21
22
namespace doris::format::parquet {
23
24
109
void ParquetProfile::init(RuntimeProfile* profile) {
25
109
    if (profile == nullptr) {
26
82
        return;
27
82
    }
28
29
27
    static const char* parquet_profile = "ParquetReader";
30
27
    ADD_TIMER_WITH_LEVEL(profile, parquet_profile, 1);
31
32
27
    filtered_row_groups = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RowGroupsFiltered", TUnit::UNIT,
33
27
                                                       parquet_profile, 1);
34
27
    filtered_row_groups_by_min_max = ADD_CHILD_COUNTER_WITH_LEVEL(
35
27
            profile, "RowGroupsFilteredByMinMax", TUnit::UNIT, parquet_profile, 1);
36
27
    filtered_row_groups_by_dictionary = ADD_CHILD_COUNTER_WITH_LEVEL(
37
27
            profile, "RowGroupsFilteredByDictionary", TUnit::UNIT, parquet_profile, 1);
38
27
    filtered_row_groups_by_bloom_filter = ADD_CHILD_COUNTER_WITH_LEVEL(
39
27
            profile, "RowGroupsFilteredByBloomFilter", TUnit::UNIT, parquet_profile, 1);
40
27
    filtered_row_groups_by_page_index = ADD_CHILD_COUNTER_WITH_LEVEL(
41
27
            profile, "RowGroupsFilteredByPageIndex", TUnit::UNIT, parquet_profile, 1);
42
27
    to_read_row_groups = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RowGroupsReadNum", TUnit::UNIT,
43
27
                                                      parquet_profile, 1);
44
27
    total_row_groups = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RowGroupsTotalNum", TUnit::UNIT,
45
27
                                                    parquet_profile, 1);
46
27
    selected_row_ranges = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "SelectedRowRanges", TUnit::UNIT,
47
27
                                                       parquet_profile, 1);
48
27
    filtered_group_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "FilteredRowsByGroup", TUnit::UNIT,
49
27
                                                       parquet_profile, 1);
50
27
    filtered_page_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "FilteredRowsByPage", TUnit::UNIT,
51
27
                                                      parquet_profile, 1);
52
27
    pages_skipped_by_data_page_filter = ADD_CHILD_COUNTER_WITH_LEVEL(
53
27
            profile, "PagesSkippedByDataPageFilter", TUnit::UNIT, parquet_profile, 1);
54
27
    data_page_filter_skip_bytes = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "DataPageFilterSkipBytes",
55
27
                                                               TUnit::BYTES, parquet_profile, 1);
56
27
    selected_rows =
57
27
            ADD_CHILD_COUNTER_WITH_LEVEL(profile, "SelectedRows", TUnit::UNIT, parquet_profile, 1);
58
27
    rows_filtered_by_conjunct = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RowsFilteredByConjunct",
59
27
                                                             TUnit::UNIT, parquet_profile, 1);
60
27
    total_batches =
61
27
            ADD_CHILD_COUNTER_WITH_LEVEL(profile, "TotalBatches", TUnit::UNIT, parquet_profile, 1);
62
27
    empty_selection_batches = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "EmptySelectionBatches",
63
27
                                                           TUnit::UNIT, parquet_profile, 1);
64
27
    range_gap_skipped_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RangeGapSkippedRows",
65
27
                                                          TUnit::UNIT, parquet_profile, 1);
66
27
    reader_read_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "ReaderReadRows", TUnit::UNIT,
67
27
                                                    parquet_profile, 1);
68
27
    reader_skip_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "ReaderSkipRows", TUnit::UNIT,
69
27
                                                    parquet_profile, 1);
70
27
    reader_select_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "ReaderSelectRows", TUnit::UNIT,
71
27
                                                      parquet_profile, 1);
72
27
    arrow_read_records_time =
73
27
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "ArrowReadRecordsTime", parquet_profile, 1);
74
27
    materialization_time =
75
27
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "MaterializationTime", parquet_profile, 1);
76
27
    lazy_read_filtered_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "FilteredRowsByLazyRead",
77
27
                                                           TUnit::UNIT, parquet_profile, 1);
78
27
    filtered_bytes = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "FilteredBytes", TUnit::BYTES,
79
27
                                                  parquet_profile, 1);
80
27
    raw_rows_read =
81
27
            ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RawRowsRead", TUnit::UNIT, parquet_profile, 1);
82
27
    column_read_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "ColumnReadTime", parquet_profile, 1);
83
27
    parse_meta_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "ParseMetaTime", parquet_profile, 1);
84
27
    parse_footer_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "ParseFooterTime", parquet_profile, 1);
85
27
    file_reader_create_time =
86
27
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "FileReaderCreateTime", parquet_profile, 1);
87
27
    open_file_num =
88
27
            ADD_CHILD_COUNTER_WITH_LEVEL(profile, "FileNum", TUnit::UNIT, parquet_profile, 1);
89
27
    page_index_read_calls = ADD_COUNTER_WITH_LEVEL(profile, "PageIndexReadCalls", TUnit::UNIT, 1);
90
27
    page_index_filter_time =
91
27
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PageIndexFilterTime", parquet_profile, 1);
92
27
    read_page_index_time =
93
27
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PageIndexReadTime", parquet_profile, 1);
94
27
    parse_page_index_time =
95
27
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PageIndexParseTime", parquet_profile, 1);
96
27
    expr_zonemap_unusable = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "ExprZoneMapUnusableEvals",
97
27
                                                         TUnit::UNIT, parquet_profile, 1);
98
27
    in_zonemap_point_check = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "InZoneMapPointCheckCount",
99
27
                                                          TUnit::UNIT, parquet_profile, 1);
100
27
    in_zonemap_range_only = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "InZoneMapRangeOnlyCount",
101
27
                                                         TUnit::UNIT, parquet_profile, 1);
102
27
    row_group_filter_time =
103
27
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "RowGroupFilterTime", parquet_profile, 1);
104
27
    file_footer_read_calls = ADD_COUNTER_WITH_LEVEL(profile, "FileFooterReadCalls", TUnit::UNIT, 1);
105
27
    file_footer_hit_cache = ADD_COUNTER_WITH_LEVEL(profile, "FileFooterHitCache", TUnit::UNIT, 1);
106
27
    decompress_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "DecompressTime", parquet_profile, 1);
107
27
    decompress_cnt = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "DecompressCount", TUnit::UNIT,
108
27
                                                  parquet_profile, 1);
109
27
    page_read_counter =
110
27
            ADD_CHILD_COUNTER_WITH_LEVEL(profile, "PageReadCount", TUnit::UNIT, parquet_profile, 1);
111
27
    page_cache_write_counter = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "PageCacheWriteCount",
112
27
                                                            TUnit::UNIT, parquet_profile, 1);
113
27
    page_cache_compressed_write_counter = ADD_CHILD_COUNTER_WITH_LEVEL(
114
27
            profile, "PageCacheCompressedWriteCount", TUnit::UNIT, parquet_profile, 1);
115
27
    page_cache_decompressed_write_counter = ADD_CHILD_COUNTER_WITH_LEVEL(
116
27
            profile, "PageCacheDecompressedWriteCount", TUnit::UNIT, parquet_profile, 1);
117
27
    page_cache_hit_counter = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "PageCacheHitCount", TUnit::UNIT,
118
27
                                                          parquet_profile, 1);
119
27
    page_cache_missing_counter = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "PageCacheMissingCount",
120
27
                                                              TUnit::UNIT, parquet_profile, 1);
121
27
    page_cache_compressed_hit_counter = ADD_CHILD_COUNTER_WITH_LEVEL(
122
27
            profile, "PageCacheCompressedHitCount", TUnit::UNIT, parquet_profile, 1);
123
27
    page_cache_decompressed_hit_counter = ADD_CHILD_COUNTER_WITH_LEVEL(
124
27
            profile, "PageCacheDecompressedHitCount", TUnit::UNIT, parquet_profile, 1);
125
27
    decode_header_time =
126
27
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PageHeaderDecodeTime", parquet_profile, 1);
127
27
    read_page_header_time =
128
27
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PageHeaderReadTime", parquet_profile, 1);
129
27
    decode_value_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "DecodeValueTime", parquet_profile, 1);
130
27
    decode_dict_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "DecodeDictTime", parquet_profile, 1);
131
27
    decode_level_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "DecodeLevelTime", parquet_profile, 1);
132
27
    decode_null_map_time =
133
27
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "DecodeNullMapTime", parquet_profile, 1);
134
27
    skip_page_header_num = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "SkipPageHeaderNum", TUnit::UNIT,
135
27
                                                        parquet_profile, 1);
136
27
    parse_page_header_num = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "ParsePageHeaderNum", TUnit::UNIT,
137
27
                                                         parquet_profile, 1);
138
27
    predicate_filter_time =
139
27
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PredicateFilterTime", parquet_profile, 1);
140
27
    dict_filter_rewrite_time =
141
27
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "DictFilterRewriteTime", parquet_profile, 1);
142
27
    convert_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "ConvertTime", parquet_profile, 1);
143
27
    bloom_filter_read_time =
144
27
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "BloomFilterReadTime", parquet_profile, 1);
145
27
}
146
147
27
void ParquetProfile::update_pruning_stats(const ParquetPruningStats& pruning_stats) const {
148
27
    COUNTER_UPDATE(filtered_row_groups,
149
27
                   pruning_stats.total_row_groups - pruning_stats.selected_row_groups);
150
27
    COUNTER_UPDATE(filtered_row_groups_by_min_max, pruning_stats.filtered_row_groups_by_statistics);
151
27
    COUNTER_UPDATE(filtered_row_groups_by_dictionary,
152
27
                   pruning_stats.filtered_row_groups_by_dictionary);
153
27
    COUNTER_UPDATE(filtered_row_groups_by_bloom_filter,
154
27
                   pruning_stats.filtered_row_groups_by_bloom_filter);
155
27
    COUNTER_UPDATE(filtered_row_groups_by_page_index,
156
27
                   pruning_stats.filtered_row_groups_by_page_index);
157
27
    COUNTER_UPDATE(to_read_row_groups, pruning_stats.selected_row_groups);
158
27
    COUNTER_UPDATE(total_row_groups, pruning_stats.total_row_groups);
159
27
    COUNTER_UPDATE(selected_row_ranges, pruning_stats.selected_row_ranges);
160
27
    COUNTER_UPDATE(filtered_group_rows, pruning_stats.filtered_group_rows);
161
27
    COUNTER_UPDATE(filtered_page_rows, pruning_stats.filtered_page_rows);
162
27
    COUNTER_UPDATE(page_index_read_calls, pruning_stats.page_index_read_calls);
163
27
    COUNTER_UPDATE(bloom_filter_read_time, pruning_stats.bloom_filter_read_time);
164
27
    COUNTER_UPDATE(row_group_filter_time, pruning_stats.row_group_filter_time);
165
27
    COUNTER_UPDATE(page_index_filter_time, pruning_stats.page_index_filter_time);
166
27
    COUNTER_UPDATE(read_page_index_time, pruning_stats.read_page_index_time);
167
27
    COUNTER_UPDATE(expr_zonemap_unusable, pruning_stats.expr_zonemap_unusable_evals);
168
27
    COUNTER_UPDATE(in_zonemap_point_check, pruning_stats.in_zonemap_point_check_count);
169
27
    COUNTER_UPDATE(in_zonemap_range_only, pruning_stats.in_zonemap_range_only_count);
170
27
}
171
172
110
ParquetPageSkipProfile ParquetProfile::page_skip_profile() const {
173
110
    return {
174
110
            .skipped_pages = pages_skipped_by_data_page_filter,
175
110
            .skipped_bytes = data_page_filter_skip_bytes,
176
110
    };
177
110
}
178
179
110
ParquetColumnReaderProfile ParquetProfile::column_reader_profile() const {
180
110
    return {
181
110
            .reader_read_rows = reader_read_rows,
182
110
            .reader_skip_rows = reader_skip_rows,
183
110
            .reader_select_rows = reader_select_rows,
184
110
            .arrow_read_records_time = arrow_read_records_time,
185
110
            .materialization_time = materialization_time,
186
110
    };
187
110
}
188
189
110
ParquetScanProfile ParquetProfile::scan_profile() const {
190
110
    return {
191
110
            .raw_rows_read = raw_rows_read,
192
110
            .selected_rows = selected_rows,
193
110
            .rows_filtered_by_conjunct = rows_filtered_by_conjunct,
194
110
            .lazy_read_filtered_rows = lazy_read_filtered_rows,
195
110
            .total_batches = total_batches,
196
110
            .empty_selection_batches = empty_selection_batches,
197
110
            .range_gap_skipped_rows = range_gap_skipped_rows,
198
110
            .column_read_time = column_read_time,
199
110
            .predicate_filter_time = predicate_filter_time,
200
110
            .column_reader_profile = column_reader_profile(),
201
110
    };
202
110
}
203
204
} // namespace doris::format::parquet