Coverage Report

Created: 2026-07-07 22:17

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
220
void ParquetProfile::init(RuntimeProfile* profile) {
25
220
    if (profile == nullptr) {
26
164
        return;
27
164
    }
28
29
56
    static const char* parquet_profile = "ParquetReader";
30
56
    ADD_TIMER_WITH_LEVEL(profile, parquet_profile, 1);
31
32
56
    filtered_row_groups = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RowGroupsFiltered", TUnit::UNIT,
33
56
                                                       parquet_profile, 1);
34
56
    filtered_row_groups_by_min_max = ADD_CHILD_COUNTER_WITH_LEVEL(
35
56
            profile, "RowGroupsFilteredByMinMax", TUnit::UNIT, parquet_profile, 1);
36
56
    filtered_row_groups_by_dictionary = ADD_CHILD_COUNTER_WITH_LEVEL(
37
56
            profile, "RowGroupsFilteredByDictionary", TUnit::UNIT, parquet_profile, 1);
38
56
    filtered_row_groups_by_bloom_filter = ADD_CHILD_COUNTER_WITH_LEVEL(
39
56
            profile, "RowGroupsFilteredByBloomFilter", TUnit::UNIT, parquet_profile, 1);
40
56
    filtered_row_groups_by_page_index = ADD_CHILD_COUNTER_WITH_LEVEL(
41
56
            profile, "RowGroupsFilteredByPageIndex", TUnit::UNIT, parquet_profile, 1);
42
56
    to_read_row_groups = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RowGroupsReadNum", TUnit::UNIT,
43
56
                                                      parquet_profile, 1);
44
56
    total_row_groups = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RowGroupsTotalNum", TUnit::UNIT,
45
56
                                                    parquet_profile, 1);
46
56
    selected_row_ranges = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "SelectedRowRanges", TUnit::UNIT,
47
56
                                                       parquet_profile, 1);
48
56
    filtered_group_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "FilteredRowsByGroup", TUnit::UNIT,
49
56
                                                       parquet_profile, 1);
50
56
    filtered_page_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "FilteredRowsByPage", TUnit::UNIT,
51
56
                                                      parquet_profile, 1);
52
56
    pages_skipped_by_data_page_filter = ADD_CHILD_COUNTER_WITH_LEVEL(
53
56
            profile, "PagesSkippedByDataPageFilter", TUnit::UNIT, parquet_profile, 1);
54
56
    data_page_filter_skip_bytes = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "DataPageFilterSkipBytes",
55
56
                                                               TUnit::BYTES, parquet_profile, 1);
56
56
    selected_rows =
57
56
            ADD_CHILD_COUNTER_WITH_LEVEL(profile, "SelectedRows", TUnit::UNIT, parquet_profile, 1);
58
56
    rows_filtered_by_conjunct = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RowsFilteredByConjunct",
59
56
                                                             TUnit::UNIT, parquet_profile, 1);
60
56
    total_batches =
61
56
            ADD_CHILD_COUNTER_WITH_LEVEL(profile, "TotalBatches", TUnit::UNIT, parquet_profile, 1);
62
56
    empty_selection_batches = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "EmptySelectionBatches",
63
56
                                                           TUnit::UNIT, parquet_profile, 1);
64
56
    range_gap_skipped_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RangeGapSkippedRows",
65
56
                                                          TUnit::UNIT, parquet_profile, 1);
66
56
    reader_read_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "ReaderReadRows", TUnit::UNIT,
67
56
                                                    parquet_profile, 1);
68
56
    reader_skip_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "ReaderSkipRows", TUnit::UNIT,
69
56
                                                    parquet_profile, 1);
70
56
    reader_select_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "ReaderSelectRows", TUnit::UNIT,
71
56
                                                      parquet_profile, 1);
72
56
    arrow_read_records_time =
73
56
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "ArrowReadRecordsTime", parquet_profile, 1);
74
56
    materialization_time =
75
56
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "MaterializationTime", parquet_profile, 1);
76
56
    lazy_read_filtered_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "FilteredRowsByLazyRead",
77
56
                                                           TUnit::UNIT, parquet_profile, 1);
78
56
    filtered_bytes = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "FilteredBytes", TUnit::BYTES,
79
56
                                                  parquet_profile, 1);
80
56
    raw_rows_read =
81
56
            ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RawRowsRead", TUnit::UNIT, parquet_profile, 1);
82
56
    column_read_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "ColumnReadTime", parquet_profile, 1);
83
56
    parse_meta_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "ParseMetaTime", parquet_profile, 1);
84
56
    parse_footer_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "ParseFooterTime", parquet_profile, 1);
85
56
    file_reader_create_time =
86
56
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "FileReaderCreateTime", parquet_profile, 1);
87
56
    open_file_num =
88
56
            ADD_CHILD_COUNTER_WITH_LEVEL(profile, "FileNum", TUnit::UNIT, parquet_profile, 1);
89
56
    page_index_read_calls = ADD_COUNTER_WITH_LEVEL(profile, "PageIndexReadCalls", TUnit::UNIT, 1);
90
56
    page_index_filter_time =
91
56
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PageIndexFilterTime", parquet_profile, 1);
92
56
    read_page_index_time =
93
56
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PageIndexReadTime", parquet_profile, 1);
94
56
    parse_page_index_time =
95
56
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PageIndexParseTime", parquet_profile, 1);
96
56
    expr_zonemap_unusable = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "ExprZoneMapUnusableEvals",
97
56
                                                         TUnit::UNIT, parquet_profile, 1);
98
56
    in_zonemap_point_check = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "InZoneMapPointCheckCount",
99
56
                                                          TUnit::UNIT, parquet_profile, 1);
100
56
    in_zonemap_range_only = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "InZoneMapRangeOnlyCount",
101
56
                                                         TUnit::UNIT, parquet_profile, 1);
102
56
    row_group_filter_time =
103
56
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "RowGroupFilterTime", parquet_profile, 1);
104
56
    file_footer_read_calls = ADD_COUNTER_WITH_LEVEL(profile, "FileFooterReadCalls", TUnit::UNIT, 1);
105
56
    file_footer_hit_cache = ADD_COUNTER_WITH_LEVEL(profile, "FileFooterHitCache", TUnit::UNIT, 1);
106
56
    decompress_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "DecompressTime", parquet_profile, 1);
107
56
    decompress_cnt = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "DecompressCount", TUnit::UNIT,
108
56
                                                  parquet_profile, 1);
109
56
    page_read_counter =
110
56
            ADD_CHILD_COUNTER_WITH_LEVEL(profile, "PageReadCount", TUnit::UNIT, parquet_profile, 1);
111
56
    page_cache_write_counter = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "PageCacheWriteCount",
112
56
                                                            TUnit::UNIT, parquet_profile, 1);
113
56
    page_cache_compressed_write_counter = ADD_CHILD_COUNTER_WITH_LEVEL(
114
56
            profile, "PageCacheCompressedWriteCount", TUnit::UNIT, parquet_profile, 1);
115
56
    page_cache_decompressed_write_counter = ADD_CHILD_COUNTER_WITH_LEVEL(
116
56
            profile, "PageCacheDecompressedWriteCount", TUnit::UNIT, parquet_profile, 1);
117
56
    page_cache_hit_counter = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "PageCacheHitCount", TUnit::UNIT,
118
56
                                                          parquet_profile, 1);
119
56
    page_cache_missing_counter = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "PageCacheMissingCount",
120
56
                                                              TUnit::UNIT, parquet_profile, 1);
121
56
    page_cache_compressed_hit_counter = ADD_CHILD_COUNTER_WITH_LEVEL(
122
56
            profile, "PageCacheCompressedHitCount", TUnit::UNIT, parquet_profile, 1);
123
56
    page_cache_decompressed_hit_counter = ADD_CHILD_COUNTER_WITH_LEVEL(
124
56
            profile, "PageCacheDecompressedHitCount", TUnit::UNIT, parquet_profile, 1);
125
56
    decode_header_time =
126
56
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PageHeaderDecodeTime", parquet_profile, 1);
127
56
    read_page_header_time =
128
56
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PageHeaderReadTime", parquet_profile, 1);
129
56
    decode_value_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "DecodeValueTime", parquet_profile, 1);
130
56
    decode_dict_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "DecodeDictTime", parquet_profile, 1);
131
56
    decode_level_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "DecodeLevelTime", parquet_profile, 1);
132
56
    decode_null_map_time =
133
56
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "DecodeNullMapTime", parquet_profile, 1);
134
56
    skip_page_header_num = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "SkipPageHeaderNum", TUnit::UNIT,
135
56
                                                        parquet_profile, 1);
136
56
    parse_page_header_num = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "ParsePageHeaderNum", TUnit::UNIT,
137
56
                                                         parquet_profile, 1);
138
56
    predicate_filter_time =
139
56
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PredicateFilterTime", parquet_profile, 1);
140
56
    dict_filter_rewrite_time =
141
56
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "DictFilterRewriteTime", parquet_profile, 1);
142
56
    convert_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "ConvertTime", parquet_profile, 1);
143
56
    bloom_filter_read_time =
144
56
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "BloomFilterReadTime", parquet_profile, 1);
145
56
}
146
147
56
void ParquetProfile::update_pruning_stats(const ParquetPruningStats& pruning_stats) const {
148
56
    COUNTER_UPDATE(filtered_row_groups,
149
56
                   pruning_stats.total_row_groups - pruning_stats.selected_row_groups);
150
56
    COUNTER_UPDATE(filtered_row_groups_by_min_max, pruning_stats.filtered_row_groups_by_statistics);
151
56
    COUNTER_UPDATE(filtered_row_groups_by_dictionary,
152
56
                   pruning_stats.filtered_row_groups_by_dictionary);
153
56
    COUNTER_UPDATE(filtered_row_groups_by_bloom_filter,
154
56
                   pruning_stats.filtered_row_groups_by_bloom_filter);
155
56
    COUNTER_UPDATE(filtered_row_groups_by_page_index,
156
56
                   pruning_stats.filtered_row_groups_by_page_index);
157
56
    COUNTER_UPDATE(to_read_row_groups, pruning_stats.selected_row_groups);
158
56
    COUNTER_UPDATE(total_row_groups, pruning_stats.total_row_groups);
159
56
    COUNTER_UPDATE(selected_row_ranges, pruning_stats.selected_row_ranges);
160
56
    COUNTER_UPDATE(filtered_group_rows, pruning_stats.filtered_group_rows);
161
56
    COUNTER_UPDATE(filtered_page_rows, pruning_stats.filtered_page_rows);
162
56
    COUNTER_UPDATE(page_index_read_calls, pruning_stats.page_index_read_calls);
163
56
    COUNTER_UPDATE(bloom_filter_read_time, pruning_stats.bloom_filter_read_time);
164
56
    COUNTER_UPDATE(row_group_filter_time, pruning_stats.row_group_filter_time);
165
56
    COUNTER_UPDATE(page_index_filter_time, pruning_stats.page_index_filter_time);
166
56
    COUNTER_UPDATE(read_page_index_time, pruning_stats.read_page_index_time);
167
56
    COUNTER_UPDATE(expr_zonemap_unusable, pruning_stats.expr_zonemap_unusable_evals);
168
56
    COUNTER_UPDATE(in_zonemap_point_check, pruning_stats.in_zonemap_point_check_count);
169
56
    COUNTER_UPDATE(in_zonemap_range_only, pruning_stats.in_zonemap_range_only_count);
170
56
}
171
172
222
ParquetPageSkipProfile ParquetProfile::page_skip_profile() const {
173
222
    return {
174
222
            .skipped_pages = pages_skipped_by_data_page_filter,
175
222
            .skipped_bytes = data_page_filter_skip_bytes,
176
222
    };
177
222
}
178
179
222
ParquetColumnReaderProfile ParquetProfile::column_reader_profile() const {
180
222
    return {
181
222
            .reader_read_rows = reader_read_rows,
182
222
            .reader_skip_rows = reader_skip_rows,
183
222
            .reader_select_rows = reader_select_rows,
184
222
            .arrow_read_records_time = arrow_read_records_time,
185
222
            .materialization_time = materialization_time,
186
222
    };
187
222
}
188
189
222
ParquetScanProfile ParquetProfile::scan_profile() const {
190
222
    return {
191
222
            .raw_rows_read = raw_rows_read,
192
222
            .selected_rows = selected_rows,
193
222
            .rows_filtered_by_conjunct = rows_filtered_by_conjunct,
194
222
            .lazy_read_filtered_rows = lazy_read_filtered_rows,
195
222
            .total_batches = total_batches,
196
222
            .empty_selection_batches = empty_selection_batches,
197
222
            .range_gap_skipped_rows = range_gap_skipped_rows,
198
222
            .column_read_time = column_read_time,
199
222
            .predicate_filter_time = predicate_filter_time,
200
222
            .column_reader_profile = column_reader_profile(),
201
222
    };
202
222
}
203
204
} // namespace doris::format::parquet