Coverage Report

Created: 2026-07-14 11:38

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
165
void ParquetProfile::init(RuntimeProfile* profile) {
25
165
    if (profile == nullptr) {
26
88
        return;
27
88
    }
28
29
77
    static const char* parquet_profile = "ParquetReader";
30
77
    ADD_TIMER_WITH_LEVEL(profile, parquet_profile, 1);
31
32
77
    filtered_row_groups = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RowGroupsFiltered", TUnit::UNIT,
33
77
                                                       parquet_profile, 1);
34
77
    filtered_row_groups_by_min_max = ADD_CHILD_COUNTER_WITH_LEVEL(
35
77
            profile, "RowGroupsFilteredByMinMax", TUnit::UNIT, parquet_profile, 1);
36
77
    filtered_row_groups_by_dictionary = ADD_CHILD_COUNTER_WITH_LEVEL(
37
77
            profile, "RowGroupsFilteredByDictionary", TUnit::UNIT, parquet_profile, 1);
38
77
    filtered_row_groups_by_bloom_filter = ADD_CHILD_COUNTER_WITH_LEVEL(
39
77
            profile, "RowGroupsFilteredByBloomFilter", TUnit::UNIT, parquet_profile, 1);
40
77
    filtered_row_groups_by_page_index = ADD_CHILD_COUNTER_WITH_LEVEL(
41
77
            profile, "RowGroupsFilteredByPageIndex", TUnit::UNIT, parquet_profile, 1);
42
77
    to_read_row_groups = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RowGroupsReadNum", TUnit::UNIT,
43
77
                                                      parquet_profile, 1);
44
77
    total_row_groups = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RowGroupsTotalNum", TUnit::UNIT,
45
77
                                                    parquet_profile, 1);
46
77
    selected_row_ranges = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "SelectedRowRanges", TUnit::UNIT,
47
77
                                                       parquet_profile, 1);
48
77
    filtered_group_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "FilteredRowsByGroup", TUnit::UNIT,
49
77
                                                       parquet_profile, 1);
50
77
    filtered_page_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "FilteredRowsByPage", TUnit::UNIT,
51
77
                                                      parquet_profile, 1);
52
77
    pages_skipped_by_data_page_filter = ADD_CHILD_COUNTER_WITH_LEVEL(
53
77
            profile, "PagesSkippedByDataPageFilter", TUnit::UNIT, parquet_profile, 1);
54
77
    data_page_filter_skip_bytes = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "DataPageFilterSkipBytes",
55
77
                                                               TUnit::BYTES, parquet_profile, 1);
56
77
    selected_rows =
57
77
            ADD_CHILD_COUNTER_WITH_LEVEL(profile, "SelectedRows", TUnit::UNIT, parquet_profile, 1);
58
77
    rows_filtered_by_conjunct = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RowsFilteredByConjunct",
59
77
                                                             TUnit::UNIT, parquet_profile, 1);
60
77
    total_batches =
61
77
            ADD_CHILD_COUNTER_WITH_LEVEL(profile, "TotalBatches", TUnit::UNIT, parquet_profile, 1);
62
77
    empty_selection_batches = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "EmptySelectionBatches",
63
77
                                                           TUnit::UNIT, parquet_profile, 1);
64
77
    range_gap_skipped_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RangeGapSkippedRows",
65
77
                                                          TUnit::UNIT, parquet_profile, 1);
66
77
    reader_read_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "ReaderReadRows", TUnit::UNIT,
67
77
                                                    parquet_profile, 1);
68
77
    reader_skip_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "ReaderSkipRows", TUnit::UNIT,
69
77
                                                    parquet_profile, 1);
70
77
    reader_select_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "ReaderSelectRows", TUnit::UNIT,
71
77
                                                      parquet_profile, 1);
72
77
    arrow_read_records_time =
73
77
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "ArrowReadRecordsTime", parquet_profile, 1);
74
77
    materialization_time =
75
77
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "MaterializationTime", parquet_profile, 1);
76
77
    lazy_read_filtered_rows = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "FilteredRowsByLazyRead",
77
77
                                                           TUnit::UNIT, parquet_profile, 1);
78
77
    filtered_bytes = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "FilteredBytes", TUnit::BYTES,
79
77
                                                  parquet_profile, 1);
80
77
    raw_rows_read =
81
77
            ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RawRowsRead", TUnit::UNIT, parquet_profile, 1);
82
77
    column_read_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "ColumnReadTime", parquet_profile, 1);
83
77
    parse_meta_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "ParseMetaTime", parquet_profile, 1);
84
77
    parse_footer_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "ParseFooterTime", parquet_profile, 1);
85
77
    file_reader_create_time =
86
77
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "FileReaderCreateTime", parquet_profile, 1);
87
77
    open_file_num =
88
77
            ADD_CHILD_COUNTER_WITH_LEVEL(profile, "FileNum", TUnit::UNIT, parquet_profile, 1);
89
77
    page_index_read_calls = ADD_COUNTER_WITH_LEVEL(profile, "PageIndexReadCalls", TUnit::UNIT, 1);
90
77
    page_index_filter_time =
91
77
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PageIndexFilterTime", parquet_profile, 1);
92
77
    read_page_index_time =
93
77
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PageIndexReadTime", parquet_profile, 1);
94
77
    parse_page_index_time =
95
77
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PageIndexParseTime", parquet_profile, 1);
96
77
    expr_zonemap_unusable = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "ExprZoneMapUnusableEvals",
97
77
                                                         TUnit::UNIT, parquet_profile, 1);
98
77
    in_zonemap_point_check = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "InZoneMapPointCheckCount",
99
77
                                                          TUnit::UNIT, parquet_profile, 1);
100
77
    in_zonemap_range_only = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "InZoneMapRangeOnlyCount",
101
77
                                                         TUnit::UNIT, parquet_profile, 1);
102
77
    row_group_filter_time =
103
77
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "RowGroupFilterTime", parquet_profile, 1);
104
77
    file_footer_read_calls = ADD_COUNTER_WITH_LEVEL(profile, "FileFooterReadCalls", TUnit::UNIT, 1);
105
77
    file_footer_hit_cache = ADD_COUNTER_WITH_LEVEL(profile, "FileFooterHitCache", TUnit::UNIT, 1);
106
77
    decompress_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "DecompressTime", parquet_profile, 1);
107
77
    decompress_cnt = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "DecompressCount", TUnit::UNIT,
108
77
                                                  parquet_profile, 1);
109
77
    page_read_counter =
110
77
            ADD_CHILD_COUNTER_WITH_LEVEL(profile, "PageReadCount", TUnit::UNIT, parquet_profile, 1);
111
77
    page_cache_write_counter = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "PageCacheWriteCount",
112
77
                                                            TUnit::UNIT, parquet_profile, 1);
113
77
    page_cache_compressed_write_counter = ADD_CHILD_COUNTER_WITH_LEVEL(
114
77
            profile, "PageCacheCompressedWriteCount", TUnit::UNIT, parquet_profile, 1);
115
77
    page_cache_decompressed_write_counter = ADD_CHILD_COUNTER_WITH_LEVEL(
116
77
            profile, "PageCacheDecompressedWriteCount", TUnit::UNIT, parquet_profile, 1);
117
77
    page_cache_hit_counter = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "PageCacheHitCount", TUnit::UNIT,
118
77
                                                          parquet_profile, 1);
119
77
    page_cache_missing_counter = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "PageCacheMissingCount",
120
77
                                                              TUnit::UNIT, parquet_profile, 1);
121
77
    page_cache_compressed_hit_counter = ADD_CHILD_COUNTER_WITH_LEVEL(
122
77
            profile, "PageCacheCompressedHitCount", TUnit::UNIT, parquet_profile, 1);
123
77
    page_cache_decompressed_hit_counter = ADD_CHILD_COUNTER_WITH_LEVEL(
124
77
            profile, "PageCacheDecompressedHitCount", TUnit::UNIT, parquet_profile, 1);
125
77
    decode_header_time =
126
77
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PageHeaderDecodeTime", parquet_profile, 1);
127
77
    read_page_header_time =
128
77
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PageHeaderReadTime", parquet_profile, 1);
129
77
    decode_value_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "DecodeValueTime", parquet_profile, 1);
130
77
    decode_dict_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "DecodeDictTime", parquet_profile, 1);
131
77
    decode_level_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "DecodeLevelTime", parquet_profile, 1);
132
77
    decode_null_map_time =
133
77
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "DecodeNullMapTime", parquet_profile, 1);
134
77
    skip_page_header_num = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "SkipPageHeaderNum", TUnit::UNIT,
135
77
                                                        parquet_profile, 1);
136
77
    parse_page_header_num = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "ParsePageHeaderNum", TUnit::UNIT,
137
77
                                                         parquet_profile, 1);
138
77
    predicate_filter_time =
139
77
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "PredicateFilterTime", parquet_profile, 1);
140
77
    dict_filter_rewrite_time =
141
77
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "DictFilterRewriteTime", parquet_profile, 1);
142
77
    dict_filter_expr_rewrite_time =
143
77
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "DictFilterExprRewriteTime", parquet_profile, 1);
144
77
    dict_filter_read_dict_time =
145
77
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "DictFilterReadDictTime", parquet_profile, 1);
146
77
    dict_filter_build_time =
147
77
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "DictFilterBuildTime", parquet_profile, 1);
148
77
    dict_filter_candidate_columns = ADD_CHILD_COUNTER_WITH_LEVEL(
149
77
            profile, "DictFilterCandidateColumns", TUnit::UNIT, parquet_profile, 1);
150
77
    dict_filter_columns = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "DictFilterColumns", TUnit::UNIT,
151
77
                                                       parquet_profile, 1);
152
77
    dict_filter_unsupported_columns = ADD_CHILD_COUNTER_WITH_LEVEL(
153
77
            profile, "DictFilterUnsupportedColumns", TUnit::UNIT, parquet_profile, 1);
154
77
    dict_filter_read_failures = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "DictFilterReadFailures",
155
77
                                                             TUnit::UNIT, parquet_profile, 1);
156
77
    rows_filtered_by_dict_filter = ADD_CHILD_COUNTER_WITH_LEVEL(profile, "RowsFilteredByDictFilter",
157
77
                                                                TUnit::UNIT, parquet_profile, 1);
158
77
    convert_time = ADD_CHILD_TIMER_WITH_LEVEL(profile, "ConvertTime", parquet_profile, 1);
159
77
    bloom_filter_read_time =
160
77
            ADD_CHILD_TIMER_WITH_LEVEL(profile, "BloomFilterReadTime", parquet_profile, 1);
161
77
}
162
163
77
void ParquetProfile::update_pruning_stats(const ParquetPruningStats& pruning_stats) const {
164
77
    COUNTER_UPDATE(filtered_row_groups,
165
77
                   pruning_stats.total_row_groups - pruning_stats.selected_row_groups);
166
77
    COUNTER_UPDATE(filtered_row_groups_by_min_max, pruning_stats.filtered_row_groups_by_statistics);
167
77
    COUNTER_UPDATE(filtered_row_groups_by_dictionary,
168
77
                   pruning_stats.filtered_row_groups_by_dictionary);
169
77
    COUNTER_UPDATE(filtered_row_groups_by_bloom_filter,
170
77
                   pruning_stats.filtered_row_groups_by_bloom_filter);
171
77
    COUNTER_UPDATE(filtered_row_groups_by_page_index,
172
77
                   pruning_stats.filtered_row_groups_by_page_index);
173
77
    COUNTER_UPDATE(to_read_row_groups, pruning_stats.selected_row_groups);
174
77
    COUNTER_UPDATE(total_row_groups, pruning_stats.total_row_groups);
175
77
    COUNTER_UPDATE(selected_row_ranges, pruning_stats.selected_row_ranges);
176
77
    COUNTER_UPDATE(filtered_group_rows, pruning_stats.filtered_group_rows);
177
77
    COUNTER_UPDATE(filtered_page_rows, pruning_stats.filtered_page_rows);
178
77
    COUNTER_UPDATE(page_index_read_calls, pruning_stats.page_index_read_calls);
179
77
    COUNTER_UPDATE(bloom_filter_read_time, pruning_stats.bloom_filter_read_time);
180
77
    COUNTER_UPDATE(row_group_filter_time, pruning_stats.row_group_filter_time);
181
77
    COUNTER_UPDATE(page_index_filter_time, pruning_stats.page_index_filter_time);
182
77
    COUNTER_UPDATE(read_page_index_time, pruning_stats.read_page_index_time);
183
77
    COUNTER_UPDATE(expr_zonemap_unusable, pruning_stats.expr_zonemap_unusable_evals);
184
77
    COUNTER_UPDATE(in_zonemap_point_check, pruning_stats.in_zonemap_point_check_count);
185
77
    COUNTER_UPDATE(in_zonemap_range_only, pruning_stats.in_zonemap_range_only_count);
186
77
}
187
188
168
ParquetPageSkipProfile ParquetProfile::page_skip_profile() const {
189
168
    return {
190
168
            .skipped_pages = pages_skipped_by_data_page_filter,
191
168
            .skipped_bytes = data_page_filter_skip_bytes,
192
168
    };
193
168
}
194
195
168
ParquetColumnReaderProfile ParquetProfile::column_reader_profile() const {
196
168
    return {
197
168
            .reader_read_rows = reader_read_rows,
198
168
            .reader_skip_rows = reader_skip_rows,
199
168
            .reader_select_rows = reader_select_rows,
200
168
            .arrow_read_records_time = arrow_read_records_time,
201
168
            .materialization_time = materialization_time,
202
168
    };
203
168
}
204
205
168
ParquetScanProfile ParquetProfile::scan_profile() const {
206
168
    return {
207
168
            .raw_rows_read = raw_rows_read,
208
168
            .selected_rows = selected_rows,
209
168
            .rows_filtered_by_conjunct = rows_filtered_by_conjunct,
210
168
            .lazy_read_filtered_rows = lazy_read_filtered_rows,
211
168
            .total_batches = total_batches,
212
168
            .empty_selection_batches = empty_selection_batches,
213
168
            .range_gap_skipped_rows = range_gap_skipped_rows,
214
168
            .column_read_time = column_read_time,
215
168
            .predicate_filter_time = predicate_filter_time,
216
168
            .dict_filter_rewrite_time = dict_filter_rewrite_time,
217
168
            .dict_filter_expr_rewrite_time = dict_filter_expr_rewrite_time,
218
168
            .dict_filter_read_dict_time = dict_filter_read_dict_time,
219
168
            .dict_filter_build_time = dict_filter_build_time,
220
168
            .dict_filter_candidate_columns = dict_filter_candidate_columns,
221
168
            .dict_filter_columns = dict_filter_columns,
222
168
            .dict_filter_unsupported_columns = dict_filter_unsupported_columns,
223
168
            .dict_filter_read_failures = dict_filter_read_failures,
224
168
            .rows_filtered_by_dict_filter = rows_filtered_by_dict_filter,
225
168
            .column_reader_profile = column_reader_profile(),
226
168
    };
227
168
}
228
229
} // namespace doris::format::parquet