thevar1able/ClickHouse
1
0
Fork 0
mirror of https://github.com/ClickHouse/ClickHouse.git synced 2024-11-21 15:12:02 +00:00
Code Issues Packages Projects Releases Wiki Activity

Updated hash dictionary nullable attribute implementation

Browse Source
This commit is contained in:
Maksim Kita 2021-03-26 21:01:56 +03:00
parent 5fd575977a
commit 3f273ef983
4 changed files with 149 additions and 157 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

91
src/Dictionaries/HashedDictionary.cpp
View File

@ -2,6 +2,8 @@
#include <ext/size.h>
#include <absl/container/flat_hash_map.h>
#include <Core/Defines.h>
#include <DataTypes/DataTypesDecimal.h>
#include <Columns/ColumnsNumber.h>
@ -155,37 +157,17 @@ ColumnPtr HashedDictionary<dictionary_key_type, sparse>::getColumn(
const DictionaryAttribute & dictionary_attribute = dict_struct.getAttribute(attribute_name, result_type);
DefaultValueProvider default_value_provider(dictionary_attribute.null_value, default_values_column);
/// TODO: Check extractor performance
DictionaryKeysExtractor<dictionary_key_type> extractor(key_columns, temporary_complex_key_arena);
const auto & requested_keys = extractor.getKeys();
auto result_column = dictionary_attribute.type->createColumn();
auto result_column = dictionary_attribute.nested_type->createColumn();
size_t attribute_index = dict_struct.attribute_name_to_index.find(attribute_name)->second;
const auto & attribute = attributes[attribute_index];
size_t requested_keys_size = requested_keys.size();
if (unlikely(attribute.is_complex_type))
{
const auto & attribute_container = std::get<ComplexAttributeCollectionType<Field>>(attribute.container);
Field row_value_to_insert;
for (size_t requested_key_index = 0; requested_key_index < requested_keys_size; ++requested_key_index)
{
auto & requested_key = requested_keys[requested_key_index];
auto it = attribute_container.find(requested_key);
if (it != attribute_container.end())
row_value_to_insert = getValueFromCell(it);
else
row_value_to_insert = default_value_provider.getDefaultValue(requested_key_index);
result_column->insert(row_value_to_insert);
}
}
else
{
auto type_call = [&](const auto & dictionary_attribute_type)
{
using Type = std::decay_t<decltype(dictionary_attribute_type)>;
@ -247,6 +229,19 @@ ColumnPtr HashedDictionary<dictionary_key_type, sparse>::getColumn(
};
callOnDictionaryAttributeType(attribute.type, type_call);
if (attribute.is_nullable_set)
{
ColumnUInt8::MutablePtr col_null_map_to = ColumnUInt8::create(requested_keys_size, false);
ColumnUInt8::Container& vec_null_map_to = col_null_map_to->getData();
for (size_t requested_key_index = 0; requested_key_index < requested_keys_size; ++requested_key_index)
{
auto key = requested_keys[requested_key_index];
vec_null_map_to[requested_key_index] = (attribute.is_nullable_set->find(key) != nullptr);
}
result_column = ColumnNullable::create(std::move(result_column), std::move(col_null_map_to));
}
query_count.fetch_add(requested_keys.size(), std::memory_order_relaxed);
@ -260,6 +255,7 @@ ColumnUInt8::Ptr HashedDictionary<dictionary_key_type, sparse>::hasKeys(const Co
if (dictionary_key_type == DictionaryKeyType::complex)
dict_struct.validateKeyTypes(key_types);
/// TODO: Check performance of extractor
Arena complex_keys_arena;
DictionaryKeysExtractor<dictionary_key_type> extractor(key_columns, complex_keys_arena);
@ -275,14 +271,18 @@ ColumnUInt8::Ptr HashedDictionary<dictionary_key_type, sparse>::hasKeys(const Co
return result;
}
/// Contaiiner
const auto & attribute = attributes.front();
getAttributeContainer(0, [&](const auto & container)
{
for (size_t requested_key_index = 0; requested_key_index < keys_size; ++requested_key_index)
{
const auto & requested_key = keys[requested_key_index];
out[requested_key_index] = container.find(requested_key) != container.end();
if (unlikely(attribute.is_nullable_set) && !out[requested_key_index])
out[requested_key_index] = attribute.is_nullable_set->find(requested_key) != nullptr;
}
});
@ -416,26 +416,17 @@ void HashedDictionary<dictionary_key_type, sparse>::createAttributes()
for (const auto & dictionary_attribute : dict_struct.attributes)
{
bool is_complex_type = dictionary_attribute.is_nullable || dictionary_attribute.is_array;
auto type_call = [&, this](const auto & dictionary_attribute_type)
{
using Type = std::decay_t<decltype(dictionary_attribute_type)>;
using AttributeType = typename Type::AttributeType;
using ValueType = DictionaryValueType<AttributeType>;
auto is_nullable_set = dictionary_attribute.is_nullable ? std::make_optional<NullableSet>() : std::optional<NullableSet>{};
std::unique_ptr<Arena> string_arena = std::is_same_v<AttributeType, String> ? std::make_unique<Arena>() : nullptr;
if (is_complex_type)
{
Attribute attribute{dictionary_attribute.underlying_type, is_complex_type, ComplexAttributeCollectionType<Field>(), std::move(string_arena)};
Attribute attribute{dictionary_attribute.underlying_type, std::move(is_nullable_set), CollectionType<ValueType>(), std::move(string_arena)};
attributes.emplace_back(std::move(attribute));
}
else
{
Attribute attribute{dictionary_attribute.underlying_type, is_complex_type, CollectionType<ValueType>(), std::move(string_arena)};
attributes.emplace_back(std::move(attribute));
}
};
callOnDictionaryAttributeType(dictionary_attribute.underlying_type, type_call);
@ -567,6 +558,9 @@ void HashedDictionary<dictionary_key_type, sparse>::blockToAttributes(const Bloc
auto key = keys_extracted_from_block[key_index];
auto it = container.find(key);
if (attribute.is_nullable_set && (attribute.is_nullable_set->find(key) != nullptr))
continue;
if (it != container.end())
continue;
@ -575,11 +569,13 @@ void HashedDictionary<dictionary_key_type, sparse>::blockToAttributes(const Bloc
attribute_column.get(key_index, column_value_to_insert);
if constexpr (std::is_same_v<AttributeValueType, Field>)
if (attribute.is_nullable_set && column_value_to_insert.isNull())
{
container.insert({key, column_value_to_insert});
attribute.is_nullable_set->insert(key);
continue;
}
else if constexpr (std::is_same_v<AttributeValueType, StringRef>)
if constexpr (std::is_same_v<AttributeValueType, StringRef>)
{
String & value_to_insert = column_value_to_insert.get<String>();
size_t value_to_insert_size = value_to_insert.size();
@ -696,6 +692,9 @@ BlockInputStreamPtr HashedDictionary<dictionary_key_type, sparse>::getBlockInput
PaddedPODArray<HashedDictionary::KeyType> keys;
if (!attributes.empty())
{
const auto & attribute = attributes.front();
getAttributeContainer(0, [&](auto & container)
{
keys.reserve(container.size());
@ -705,7 +704,17 @@ BlockInputStreamPtr HashedDictionary<dictionary_key_type, sparse>::getBlockInput
(void)(value);
keys.emplace_back(key);
}
if (attribute.is_nullable_set)
{
const auto & is_nullable_set = *attribute.is_nullable_set;
keys.reserve(is_nullable_set.size());
for (auto & node : is_nullable_set)
keys.emplace_back(node.getKey());
}
});
}
if constexpr (dictionary_key_type == DictionaryKeyType::simple)
return std::make_shared<DictionaryBlockInputStream>(shared_from_this(), max_block_size, std::move(keys), column_names);
@ -721,13 +730,6 @@ void HashedDictionary<dictionary_key_type, sparse>::getAttributeContainer(size_t
auto & attribute = attributes[attribute_index];
if (unlikely(attribute.is_complex_type))
{
auto & attribute_container = std::get<ComplexAttributeCollectionType<Field>>(attribute.container);
std::forward<GetContainerFunc>(get_container_func)(attribute_container);
}
else
{
auto type_call = [&](const auto & dictionary_attribute_type)
{
using Type = std::decay_t<decltype(dictionary_attribute_type)>;
@ -739,7 +741,6 @@ void HashedDictionary<dictionary_key_type, sparse>::getAttributeContainer(size_t
};
callOnDictionaryAttributeType(attribute.type, type_call);
}
}
template <DictionaryKeyType dictionary_key_type, bool sparse>

21
src/Dictionaries/HashedDictionary.h
View File

@ -6,7 +6,6 @@
#include <optional>
#include <sparsehash/sparse_hash_map>
#include <absl/container/flat_hash_map.h>
#include <ext/range.h>
#include <Common/HashTable/HashMap.h>
@ -116,12 +115,6 @@ private:
HashMap<UInt64, Value>,
HashMapWithSavedHash<StringRef, Value, DefaultHash<StringRef>>>;
template <typename Value>
using ComplexAttributeCollectionTypeNonSparse = std::conditional_t<
dictionary_key_type == DictionaryKeyType::simple,
absl::flat_hash_map<UInt64, Value, DefaultHash<UInt64>>,
absl::flat_hash_map<StringRef, Value, DefaultHash<StringRef>>>;
#if !defined(ARCADIA_BUILD)
template <typename Key, typename Value>
using SparseHashMap = google::sparse_hash_map<Key, Value, DefaultHash<Key>>;
@ -139,16 +132,13 @@ private:
template <typename Value>
using CollectionType = std::conditional_t<sparse, CollectionTypeSparse<Value>, CollectionTypeNonSparse<Value>>;
template <typename Value>
using ComplexAttributeCollectionType = std::conditional_t<
sparse,
CollectionTypeSparse<Value>,
ComplexAttributeCollectionTypeNonSparse<Value>>;
using NullableSet = HashSet<KeyType, DefaultHash<KeyType>>;
struct Attribute final
{
AttributeUnderlyingType type;
bool is_complex_type;
std::optional<NullableSet> is_nullable_set;
std::variant<
CollectionType<UInt8>,
CollectionType<UInt16>,
@ -164,10 +154,11 @@ private:
CollectionType<Decimal128>,
CollectionType<Float32>,
CollectionType<Float64>,
CollectionType<StringRef>,
ComplexAttributeCollectionType<Field>>
CollectionType<StringRef>>
container;
std::unique_ptr<Arena> string_arena;
};
void createAttributes();

8
tests/queries/0_stateless/01778_hierarchical_dictionaries.reference
View File

@ -51,21 +51,21 @@ Get is in hierarchy
0
Get children
[1]
[2,3]
[3,2]
[4]
[]
[]
[]
Get all descendants
[1,2,3,4]
[2,3,4]
[1,3,2,4]
[3,2,4]
[4]
[]
[]
[]
Get descendants at first level
[1]
[2,3]
[3,2]
[4]
[]
[]

60
tests/queries/0_stateless/01778_hierarchical_dictionaries.sql
View File

@ -1,95 +1,95 @@
DROP DATABASE IF EXISTS 01778_db;
CREATE DATABASE 01778_db;
CREATE TABLE 01778_db.simple_key_hierarchy_source_table (id UInt64, parent_id UInt64) ENGINE = TinyLog;
INSERT INTO 01778_db.simple_key_hierarchy_source_table VALUES (1, 0), (2, 1), (3, 1), (4, 2);
CREATE TABLE 01778_db.hierarchy_source_table (id UInt64, parent_id UInt64) ENGINE = TinyLog;
INSERT INTO 01778_db.hierarchy_source_table VALUES (1, 0), (2, 1), (3, 1), (4, 2);
CREATE DICTIONARY 01778_db.simple_key_hierarchy_flat_dictionary
CREATE DICTIONARY 01778_db.hierarchy_flat_dictionary
(
id UInt64,
parent_id UInt64 HIERARCHICAL
)
PRIMARY KEY id
SOURCE(CLICKHOUSE(HOST 'localhost' PORT tcpPort() USER 'default' TABLE 'simple_key_hierarchy_source_table' DB '01778_db'))
SOURCE(CLICKHOUSE(HOST 'localhost' PORT tcpPort() USER 'default' TABLE 'hierarchy_source_table' DB '01778_db'))
LAYOUT(FLAT())
LIFETIME(MIN 1 MAX 1000);
SELECT 'Flat dictionary';
SELECT 'Get hierarchy';
SELECT dictGetHierarchy('01778_db.simple_key_hierarchy_flat_dictionary', number) FROM system.numbers LIMIT 6;
SELECT dictGetHierarchy('01778_db.hierarchy_flat_dictionary', number) FROM system.numbers LIMIT 6;
SELECT 'Get is in hierarchy';
SELECT dictIsIn('01778_db.simple_key_hierarchy_flat_dictionary', number, number) FROM system.numbers LIMIT 6;
SELECT dictIsIn('01778_db.hierarchy_flat_dictionary', number, number) FROM system.numbers LIMIT 6;
SELECT 'Get children';
SELECT dictGetChildren('01778_db.simple_key_hierarchy_flat_dictionary', number) FROM system.numbers LIMIT 6;
SELECT dictGetChildren('01778_db.hierarchy_flat_dictionary', number) FROM system.numbers LIMIT 6;
SELECT 'Get all descendants';
SELECT dictGetDescendants('01778_db.simple_key_hierarchy_flat_dictionary', number) FROM system.numbers LIMIT 6;
SELECT dictGetDescendants('01778_db.hierarchy_flat_dictionary', number) FROM system.numbers LIMIT 6;
SELECT 'Get descendants at first level';
SELECT dictGetDescendants('01778_db.simple_key_hierarchy_flat_dictionary', number, 1) FROM system.numbers LIMIT 6;
SELECT dictGetDescendants('01778_db.hierarchy_flat_dictionary', number, 1) FROM system.numbers LIMIT 6;
DROP DICTIONARY 01778_db.simple_key_hierarchy_flat_dictionary;
DROP DICTIONARY 01778_db.hierarchy_flat_dictionary;
CREATE DICTIONARY 01778_db.simple_key_hierarchy_hashed_dictionary
CREATE DICTIONARY 01778_db.hierarchy_hashed_dictionary
(
id UInt64,
parent_id UInt64 HIERARCHICAL
)
PRIMARY KEY id
SOURCE(CLICKHOUSE(HOST 'localhost' PORT tcpPort() USER 'default' TABLE 'simple_key_hierarchy_source_table' DB '01778_db'))
LAYOUT(FLAT())
SOURCE(CLICKHOUSE(HOST 'localhost' PORT tcpPort() USER 'default' TABLE 'hierarchy_source_table' DB '01778_db'))
LAYOUT(HASHED())
LIFETIME(MIN 1 MAX 1000);
SELECT 'Hashed dictionary';
SELECT 'Get hierarchy';
SELECT dictGetHierarchy('01778_db.simple_key_hierarchy_hashed_dictionary', number) FROM system.numbers LIMIT 6;
SELECT dictGetHierarchy('01778_db.hierarchy_hashed_dictionary', number) FROM system.numbers LIMIT 6;
SELECT 'Get is in hierarchy';
SELECT dictIsIn('01778_db.simple_key_hierarchy_hashed_dictionary', number, number) FROM system.numbers LIMIT 6;
SELECT dictIsIn('01778_db.hierarchy_hashed_dictionary', number, number) FROM system.numbers LIMIT 6;
SELECT 'Get children';
SELECT dictGetChildren('01778_db.simple_key_hierarchy_hashed_dictionary', number) FROM system.numbers LIMIT 6;
SELECT dictGetChildren('01778_db.hierarchy_hashed_dictionary', number) FROM system.numbers LIMIT 6;
SELECT 'Get all descendants';
SELECT dictGetDescendants('01778_db.simple_key_hierarchy_hashed_dictionary', number) FROM system.numbers LIMIT 6;
SELECT dictGetDescendants('01778_db.hierarchy_hashed_dictionary', number) FROM system.numbers LIMIT 6;
SELECT 'Get descendants at first level';
SELECT dictGetDescendants('01778_db.simple_key_hierarchy_hashed_dictionary', number, 1) FROM system.numbers LIMIT 6;
SELECT dictGetDescendants('01778_db.hierarchy_hashed_dictionary', number, 1) FROM system.numbers LIMIT 6;
DROP DICTIONARY 01778_db.simple_key_hierarchy_hashed_dictionary;
DROP DICTIONARY 01778_db.hierarchy_hashed_dictionary;
CREATE DICTIONARY 01778_db.simple_key_hierarchy_cache_dictionary
CREATE DICTIONARY 01778_db.hierarchy_cache_dictionary
(
id UInt64,
parent_id UInt64 HIERARCHICAL
)
PRIMARY KEY id
SOURCE(CLICKHOUSE(HOST 'localhost' PORT tcpPort() USER 'default' TABLE 'simple_key_hierarchy_source_table' DB '01778_db'))
SOURCE(CLICKHOUSE(HOST 'localhost' PORT tcpPort() USER 'default' TABLE 'hierarchy_source_table' DB '01778_db'))
LAYOUT(CACHE(SIZE_IN_CELLS 10))
LIFETIME(MIN 1 MAX 1000);
SELECT 'Cache dictionary';
SELECT 'Get hierarchy';
SELECT dictGetHierarchy('01778_db.simple_key_hierarchy_cache_dictionary', number) FROM system.numbers LIMIT 6;
SELECT dictGetHierarchy('01778_db.hierarchy_cache_dictionary', number) FROM system.numbers LIMIT 6;
SELECT 'Get is in hierarchy';
SELECT dictIsIn('01778_db.simple_key_hierarchy_cache_dictionary', number, number) FROM system.numbers LIMIT 6;
SELECT dictIsIn('01778_db.hierarchy_cache_dictionary', number, number) FROM system.numbers LIMIT 6;
DROP DICTIONARY 01778_db.simple_key_hierarchy_cache_dictionary;
DROP DICTIONARY 01778_db.hierarchy_cache_dictionary;
CREATE DICTIONARY 01778_db.simple_key_hierarchy_direct_dictionary
CREATE DICTIONARY 01778_db.hierarchy_direct_dictionary
(
id UInt64,
parent_id UInt64 HIERARCHICAL
)
PRIMARY KEY id
SOURCE(CLICKHOUSE(HOST 'localhost' PORT tcpPort() USER 'default' TABLE 'simple_key_hierarchy_source_table' DB '01778_db'))
SOURCE(CLICKHOUSE(HOST 'localhost' PORT tcpPort() USER 'default' TABLE 'hierarchy_source_table' DB '01778_db'))
LAYOUT(DIRECT());
SELECT 'Direct dictionary';
SELECT 'Get hierarchy';
SELECT dictGetHierarchy('01778_db.simple_key_hierarchy_direct_dictionary', number) FROM system.numbers LIMIT 6;
SELECT dictGetHierarchy('01778_db.hierarchy_direct_dictionary', number) FROM system.numbers LIMIT 6;
SELECT 'Get is in hierarchy';
SELECT dictIsIn('01778_db.simple_key_hierarchy_direct_dictionary', number, number) FROM system.numbers LIMIT 6;
SELECT dictIsIn('01778_db.hierarchy_direct_dictionary', number, number) FROM system.numbers LIMIT 6;
DROP DICTIONARY 01778_db.simple_key_hierarchy_direct_dictionary;
DROP DICTIONARY 01778_db.hierarchy_direct_dictionary;
DROP TABLE 01778_db.simple_key_hierarchy_source_table;
DROP TABLE 01778_db.hierarchy_source_table;
DROP DATABASE 01778_db;