1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
|
#pragma once
#include <Storages/MergeTree/MergeTreeBaseSelectProcessor.h>
#include <Storages/MergeTree/MergeTreeData.h>
#include <Storages/MergeTree/MarkRange.h>
#include <Storages/MergeTree/MergeTreeBlockReadUtils.h>
#include <Storages/MergeTree/MergeTreeReadPool.h>
#include <Storages/SelectQueryInfo.h>
namespace DB
{
/// Used to read data from single part with select query
/// Cares about PREWHERE, virtual columns, indexes etc.
/// To read data from multiple parts, Storage (MergeTree) creates multiple such objects.
class MergeTreeSelectAlgorithm : public IMergeTreeSelectAlgorithm
{
public:
MergeTreeSelectAlgorithm(
const MergeTreeData & storage,
const StorageSnapshotPtr & storage_snapshot_,
const MergeTreeData::DataPartPtr & owned_data_part_,
const AlterConversionsPtr & alter_conversions_,
UInt64 max_block_size_rows,
size_t preferred_block_size_bytes,
size_t preferred_max_column_in_block_size_bytes,
Names required_columns_,
MarkRanges mark_ranges,
bool use_uncompressed_cache,
const PrewhereInfoPtr & prewhere_info,
const ExpressionActionsSettings & actions_settings_,
const MergeTreeReaderSettings & reader_settings,
MergeTreeInOrderReadPoolParallelReplicasPtr pool_,
const Names & virt_column_names = {},
size_t part_index_in_query_ = 0,
bool has_limit_below_one_block_ = false);
~MergeTreeSelectAlgorithm() override;
protected:
/// Defer initialization from constructor, because it may be heavy
/// and it's better to do it lazily in `getNewTaskImpl`, which is executing in parallel.
void initializeReaders();
void finish() final;
/// Used by Task
Names required_columns;
/// Names from header. Used in order to order columns in read blocks.
Names ordered_names;
NameSet column_name_set;
MergeTreeReadTaskColumns task_columns;
/// Data part will not be removed if the pointer owns it
MergeTreeData::DataPartPtr data_part;
/// Alter converversionss that should be applied on-fly for part.
AlterConversionsPtr alter_conversions;
/// Cache getSampleBlock call, which might be heavy.
Block sample_block;
/// Mark ranges we should read (in ascending order)
MarkRanges all_mark_ranges;
/// Value of _part_index virtual column (used only in SelectExecutor)
size_t part_index_in_query = 0;
/// If true, every task will be created only with one range.
/// It reduces amount of read data for queries with small LIMIT.
bool has_limit_below_one_block = false;
/// Pool for reading in order
MergeTreeInOrderReadPoolParallelReplicasPtr pool;
size_t total_rows = 0;
};
}
|
