add ColumnAliasesVisitor to improve partition prune for mergetree

This commit is contained in:
sundy-li 2020-11-14 14:01:34 +08:00
parent c2205498b2
commit ad99dab3ad
21 changed files with 215 additions and 23 deletions

View File

@ -0,0 +1,50 @@
#include <Interpreters/ColumnAliasesVisitor.h>
#include <Interpreters/IdentifierSemantic.h>
#include <Interpreters/addTypeConversionToAST.h>
#include <Parsers/ASTTablesInSelectQuery.h>
#include <Parsers/ASTSelectWithUnionQuery.h>
#include <Parsers/ASTSelectQuery.h>
#include <Parsers/ASTSubquery.h>
#include <Parsers/ASTAlterQuery.h>
#include <Parsers/ASTInsertQuery.h>
#include <Parsers/ASTIdentifier.h>
namespace DB
{
bool ColumnAliasesMatcher::needChildVisit(const ASTPtr & node, const ASTPtr &)
{
return !(node->as<ASTTableExpression>()
|| node->as<ASTSubquery>()
|| node->as<ASTArrayJoin>()
|| node->as<ASTSelectWithUnionQuery>());
}
void ColumnAliasesMatcher::visit(ASTPtr & ast, Data & data)
{
if (auto * node = ast->as<ASTIdentifier>())
{
if (auto column_name = IdentifierSemantic::getColumnName(*node))
{
if (const auto column_default = data.columns.getDefault(column_name.value()))
{
if (column_default.value().kind == ColumnDefaultKind::Alias)
{
const auto alias_columns = data.columns.getAliases();
for (const auto & alias_column : alias_columns)
{
if (alias_column.name == column_name.value())
{
ast = addTypeConversionToAST(column_default.value().expression->clone(), alias_column.type->getName());
//revisit ast to track recursive alias columns
Visitor(data).visit(ast);
break;
}
}
}
}
}
}
}
}

View File

@ -0,0 +1,37 @@
#pragma once
#include <Interpreters/Aliases.h>
#include <Interpreters/InDepthNodeVisitor.h>
#include <Storages/ColumnsDescription.h>
namespace DB
{
class IAST;
using ASTPtr = std::shared_ptr<IAST>;
class IDataType;
using DataTypePtr = std::shared_ptr<const IDataType>;
/// Visits AST node to rewrite alias columns in filter query
/// Currently works only in `KeyCondition` of select query
class ColumnAliasesMatcher
{
public:
using Visitor = InDepthNodeVisitor<ColumnAliasesMatcher, false>;
struct Data
{
const ColumnsDescription & columns;
Data(const ColumnsDescription & columns_)
: columns(columns_)
{}
};
static void visit(ASTPtr & ast, Data & data);
static bool needChildVisit(const ASTPtr & node, const ASTPtr & child);
};
using ColumnAliasesVisitor = ColumnAliasesMatcher::Visitor;
}

View File

@ -30,6 +30,7 @@
#include <Interpreters/JoinSwitcher.h> #include <Interpreters/JoinSwitcher.h>
#include <Interpreters/JoinedTables.h> #include <Interpreters/JoinedTables.h>
#include <Interpreters/QueryAliasesVisitor.h> #include <Interpreters/QueryAliasesVisitor.h>
#include <Interpreters/ColumnAliasesVisitor.h>
#include <Processors/Pipe.h> #include <Processors/Pipe.h>
#include <Processors/Sources/SourceFromInputStream.h> #include <Processors/Sources/SourceFromInputStream.h>
@ -1291,8 +1292,11 @@ void InterpreterSelectQuery::executeFetchColumns(
{ {
auto column_decl = storage_columns.get(column); auto column_decl = storage_columns.get(column);
/// TODO: can make CAST only if the type is different (but requires SyntaxAnalyzer). /// TODO: can make CAST only if the type is different (but requires SyntaxAnalyzer).
auto cast_column_default = addTypeConversionToAST(column_default->expression->clone(), column_decl.type->getName()); column_expr = addTypeConversionToAST(column_default->expression->clone(), column_decl.type->getName());
column_expr = setAlias(cast_column_default->clone(), column); // recursive visit for alias to alias
ColumnAliasesVisitor::Data data(storage_columns);
ColumnAliasesVisitor(data).visit(column_expr);
column_expr = setAlias(column_expr, column);
} }
else else
column_expr = std::make_shared<ASTIdentifier>(column); column_expr = std::make_shared<ASTIdentifier>(column);

View File

@ -34,6 +34,7 @@ SRCS(
ClusterProxy/SelectStreamFactory.cpp ClusterProxy/SelectStreamFactory.cpp
ClusterProxy/executeQuery.cpp ClusterProxy/executeQuery.cpp
CollectJoinOnKeysVisitor.cpp CollectJoinOnKeysVisitor.cpp
ColumnAliasesVisitor.cpp
Context.cpp Context.cpp
CrashLog.cpp CrashLog.cpp
CrossToInnerJoinVisitor.cpp CrossToInnerJoinVisitor.cpp

View File

@ -13,6 +13,7 @@
#include <Common/typeid_cast.h> #include <Common/typeid_cast.h>
#include <Interpreters/convertFieldToType.h> #include <Interpreters/convertFieldToType.h>
#include <Interpreters/Set.h> #include <Interpreters/Set.h>
#include <Interpreters/ColumnAliasesVisitor.h>
#include <Parsers/queryToString.h> #include <Parsers/queryToString.h>
#include <Parsers/ASTLiteral.h> #include <Parsers/ASTLiteral.h>
#include <Parsers/ASTSubquery.h> #include <Parsers/ASTSubquery.h>
@ -371,11 +372,12 @@ Block KeyCondition::getBlockWithConstants(
KeyCondition::KeyCondition( KeyCondition::KeyCondition(
const SelectQueryInfo & query_info, const SelectQueryInfo & query_info,
const Context & context, const Context & context,
const ColumnsDescription & columns_desc_,
const Names & key_column_names, const Names & key_column_names,
const ExpressionActionsPtr & key_expr_, const ExpressionActionsPtr & key_expr_,
bool single_point_, bool single_point_,
bool strict_) bool strict_)
: key_expr(key_expr_), prepared_sets(query_info.sets), single_point(single_point_), strict(strict_) : key_expr(key_expr_), prepared_sets(query_info.sets), columns_desc(columns_desc_), single_point(single_point_), strict(strict_)
{ {
for (size_t i = 0, size = key_column_names.size(); i < size; ++i) for (size_t i = 0, size = key_column_names.size(); i < size; ++i)
{ {
@ -405,7 +407,12 @@ KeyCondition::KeyCondition(
* To overcome the problem, before parsing the AST we transform it to its semantically equivalent form where all NOT's * To overcome the problem, before parsing the AST we transform it to its semantically equivalent form where all NOT's
* are pushed down and applied (when possible) to leaf nodes. * are pushed down and applied (when possible) to leaf nodes.
*/ */
traverseAST(cloneASTWithInversionPushDown(filter_query), context, block_with_constants);
auto cloned_query = cloneASTWithInversionPushDown(filter_query);
ColumnAliasesVisitor::Data data{columns_desc};
ColumnAliasesVisitor(data).visit(cloned_query);
traverseAST(cloned_query, context, block_with_constants);
} }
else else
{ {

View File

@ -8,6 +8,7 @@
#include <Parsers/ASTSelectQuery.h> #include <Parsers/ASTSelectQuery.h>
#include <Parsers/ASTFunction.h> #include <Parsers/ASTFunction.h>
#include <Storages/SelectQueryInfo.h> #include <Storages/SelectQueryInfo.h>
#include <Storages/ColumnsDescription.h>
namespace DB namespace DB
@ -230,6 +231,7 @@ public:
KeyCondition( KeyCondition(
const SelectQueryInfo & query_info, const SelectQueryInfo & query_info,
const Context & context, const Context & context,
const ColumnsDescription & columns_desc_,
const Names & key_column_names, const Names & key_column_names,
const ExpressionActionsPtr & key_expr, const ExpressionActionsPtr & key_expr,
bool single_point_ = false, bool single_point_ = false,
@ -418,6 +420,7 @@ private:
ColumnIndices key_columns; ColumnIndices key_columns;
ExpressionActionsPtr key_expr; ExpressionActionsPtr key_expr;
PreparedSets prepared_sets; PreparedSets prepared_sets;
const ColumnsDescription & columns_desc;
// If true, always allow key_expr to be wrapped by function // If true, always allow key_expr to be wrapped by function
bool single_point; bool single_point;

View File

@ -209,7 +209,7 @@ QueryPlanPtr MergeTreeDataSelectExecutor::readFromParts(
const auto & primary_key = metadata_snapshot->getPrimaryKey(); const auto & primary_key = metadata_snapshot->getPrimaryKey();
Names primary_key_columns = primary_key.column_names; Names primary_key_columns = primary_key.column_names;
KeyCondition key_condition(query_info, context, primary_key_columns, primary_key.expression); KeyCondition key_condition(query_info, context, metadata_snapshot->getColumns(), primary_key_columns, primary_key.expression);
if (settings.force_primary_key && key_condition.alwaysUnknownOrTrue()) if (settings.force_primary_key && key_condition.alwaysUnknownOrTrue())
{ {
@ -221,8 +221,8 @@ QueryPlanPtr MergeTreeDataSelectExecutor::readFromParts(
std::optional<PartitionPruner> partition_pruner; std::optional<PartitionPruner> partition_pruner;
if (data.minmax_idx_expr) if (data.minmax_idx_expr)
{ {
minmax_idx_condition.emplace(query_info, context, data.minmax_idx_columns, data.minmax_idx_expr); minmax_idx_condition.emplace(query_info, context, metadata_snapshot->getColumns(), data.minmax_idx_columns, data.minmax_idx_expr);
partition_pruner.emplace(metadata_snapshot->getPartitionKey(), query_info, context, false /* strict */); partition_pruner.emplace(metadata_snapshot->getPartitionKey(), metadata_snapshot->getColumns(), query_info, context, false /* strict */);
if (settings.force_index_by_date && (minmax_idx_condition->alwaysUnknownOrTrue() && partition_pruner->isUseless())) if (settings.force_index_by_date && (minmax_idx_condition->alwaysUnknownOrTrue() && partition_pruner->isUseless()))
{ {
@ -559,7 +559,7 @@ QueryPlanPtr MergeTreeDataSelectExecutor::readFromParts(
for (const auto & index : metadata_snapshot->getSecondaryIndices()) for (const auto & index : metadata_snapshot->getSecondaryIndices())
{ {
auto index_helper = MergeTreeIndexFactory::instance().get(index); auto index_helper = MergeTreeIndexFactory::instance().get(index);
auto condition = index_helper->createIndexCondition(query_info, context); auto condition = index_helper->createIndexCondition(query_info, metadata_snapshot->getColumns(), context);
if (!condition->alwaysUnknownOrTrue()) if (!condition->alwaysUnknownOrTrue())
useful_indices.emplace_back(index_helper, condition); useful_indices.emplace_back(index_helper, condition);
} }

View File

@ -67,7 +67,7 @@ MergeTreeIndexAggregatorPtr MergeTreeIndexBloomFilter::createIndexAggregator() c
return std::make_shared<MergeTreeIndexAggregatorBloomFilter>(bits_per_row, hash_functions, index.column_names); return std::make_shared<MergeTreeIndexAggregatorBloomFilter>(bits_per_row, hash_functions, index.column_names);
} }
MergeTreeIndexConditionPtr MergeTreeIndexBloomFilter::createIndexCondition(const SelectQueryInfo & query_info, const Context & context) const MergeTreeIndexConditionPtr MergeTreeIndexBloomFilter::createIndexCondition(const SelectQueryInfo & query_info, const ColumnsDescription & /*columns*/, const Context & context) const
{ {
return std::make_shared<MergeTreeIndexConditionBloomFilter>(query_info, context, index.sample_block, hash_functions); return std::make_shared<MergeTreeIndexConditionBloomFilter>(query_info, context, index.sample_block, hash_functions);
} }

View File

@ -20,7 +20,7 @@ public:
MergeTreeIndexAggregatorPtr createIndexAggregator() const override; MergeTreeIndexAggregatorPtr createIndexAggregator() const override;
MergeTreeIndexConditionPtr createIndexCondition(const SelectQueryInfo & query_info, const Context & context) const override; MergeTreeIndexConditionPtr createIndexCondition(const SelectQueryInfo & query_info, const ColumnsDescription & columns, const Context & context) const override;
bool mayBenefitFromIndexForIn(const ASTPtr & node) const override; bool mayBenefitFromIndexForIn(const ASTPtr & node) const override;

View File

@ -562,7 +562,7 @@ MergeTreeIndexAggregatorPtr MergeTreeIndexFullText::createIndexAggregator() cons
} }
MergeTreeIndexConditionPtr MergeTreeIndexFullText::createIndexCondition( MergeTreeIndexConditionPtr MergeTreeIndexFullText::createIndexCondition(
const SelectQueryInfo & query, const Context & context) const const SelectQueryInfo & query, const ColumnsDescription & /*columns*/, const Context & context) const
{ {
return std::make_shared<MergeTreeConditionFullText>(query, context, index.sample_block, params, token_extractor.get()); return std::make_shared<MergeTreeConditionFullText>(query, context, index.sample_block, params, token_extractor.get());
}; };

View File

@ -199,7 +199,7 @@ public:
MergeTreeIndexAggregatorPtr createIndexAggregator() const override; MergeTreeIndexAggregatorPtr createIndexAggregator() const override;
MergeTreeIndexConditionPtr createIndexCondition( MergeTreeIndexConditionPtr createIndexCondition(
const SelectQueryInfo & query, const Context & context) const override; const SelectQueryInfo & query, const ColumnsDescription & columns, const Context & context) const override;
bool mayBenefitFromIndexForIn(const ASTPtr & node) const override; bool mayBenefitFromIndexForIn(const ASTPtr & node) const override;

View File

@ -132,9 +132,10 @@ void MergeTreeIndexAggregatorMinMax::update(const Block & block, size_t * pos, s
MergeTreeIndexConditionMinMax::MergeTreeIndexConditionMinMax( MergeTreeIndexConditionMinMax::MergeTreeIndexConditionMinMax(
const IndexDescription & index, const IndexDescription & index,
const SelectQueryInfo & query, const SelectQueryInfo & query,
const ColumnsDescription & columns,
const Context & context) const Context & context)
: index_data_types(index.data_types) : index_data_types(index.data_types)
, condition(query, context, index.column_names, index.expression) , condition(query, context, columns, index.column_names, index.expression)
{ {
} }
@ -169,9 +170,9 @@ MergeTreeIndexAggregatorPtr MergeTreeIndexMinMax::createIndexAggregator() const
} }
MergeTreeIndexConditionPtr MergeTreeIndexMinMax::createIndexCondition( MergeTreeIndexConditionPtr MergeTreeIndexMinMax::createIndexCondition(
const SelectQueryInfo & query, const Context & context) const const SelectQueryInfo & query, const ColumnsDescription & columns, const Context & context) const
{ {
return std::make_shared<MergeTreeIndexConditionMinMax>(index, query, context); return std::make_shared<MergeTreeIndexConditionMinMax>(index, query, columns, context);
}; };
bool MergeTreeIndexMinMax::mayBenefitFromIndexForIn(const ASTPtr & node) const bool MergeTreeIndexMinMax::mayBenefitFromIndexForIn(const ASTPtr & node) const

View File

@ -52,6 +52,7 @@ public:
MergeTreeIndexConditionMinMax( MergeTreeIndexConditionMinMax(
const IndexDescription & index, const IndexDescription & index,
const SelectQueryInfo & query, const SelectQueryInfo & query,
const ColumnsDescription & columns,
const Context & context); const Context & context);
bool alwaysUnknownOrTrue() const override; bool alwaysUnknownOrTrue() const override;
@ -78,7 +79,7 @@ public:
MergeTreeIndexAggregatorPtr createIndexAggregator() const override; MergeTreeIndexAggregatorPtr createIndexAggregator() const override;
MergeTreeIndexConditionPtr createIndexCondition( MergeTreeIndexConditionPtr createIndexCondition(
const SelectQueryInfo & query, const Context & context) const override; const SelectQueryInfo & query, const ColumnsDescription & columns, const Context & context) const override;
bool mayBenefitFromIndexForIn(const ASTPtr & node) const override; bool mayBenefitFromIndexForIn(const ASTPtr & node) const override;
}; };

View File

@ -468,7 +468,7 @@ MergeTreeIndexAggregatorPtr MergeTreeIndexSet::createIndexAggregator() const
} }
MergeTreeIndexConditionPtr MergeTreeIndexSet::createIndexCondition( MergeTreeIndexConditionPtr MergeTreeIndexSet::createIndexCondition(
const SelectQueryInfo & query, const Context & context) const const SelectQueryInfo & query, const ColumnsDescription & /*columns*/, const Context & context) const
{ {
return std::make_shared<MergeTreeIndexConditionSet>(index.name, index.sample_block, max_rows, query, context); return std::make_shared<MergeTreeIndexConditionSet>(index.name, index.sample_block, max_rows, query, context);
}; };

View File

@ -129,7 +129,7 @@ public:
MergeTreeIndexAggregatorPtr createIndexAggregator() const override; MergeTreeIndexAggregatorPtr createIndexAggregator() const override;
MergeTreeIndexConditionPtr createIndexCondition( MergeTreeIndexConditionPtr createIndexCondition(
const SelectQueryInfo & query, const Context & context) const override; const SelectQueryInfo & query, const ColumnsDescription & columns, const Context & context) const override;
bool mayBenefitFromIndexForIn(const ASTPtr & node) const override; bool mayBenefitFromIndexForIn(const ASTPtr & node) const override;

View File

@ -85,7 +85,7 @@ struct IMergeTreeIndex
virtual MergeTreeIndexAggregatorPtr createIndexAggregator() const = 0; virtual MergeTreeIndexAggregatorPtr createIndexAggregator() const = 0;
virtual MergeTreeIndexConditionPtr createIndexCondition( virtual MergeTreeIndexConditionPtr createIndexCondition(
const SelectQueryInfo & query_info, const Context & context) const = 0; const SelectQueryInfo & query_info, const ColumnsDescription & columns, const Context & context) const = 0;
Names getColumnsRequiredForIndexCalc() const { return index.expression->getRequiredColumns(); } Names getColumnsRequiredForIndexCalc() const { return index.expression->getRequiredColumns(); }

View File

@ -3,6 +3,7 @@
#include <unordered_map> #include <unordered_map>
#include <Storages/KeyDescription.h> #include <Storages/KeyDescription.h>
#include <Storages/ColumnsDescription.h>
#include <Storages/MergeTree/IMergeTreeDataPart.h> #include <Storages/MergeTree/IMergeTreeDataPart.h>
#include <Storages/MergeTree/KeyCondition.h> #include <Storages/MergeTree/KeyCondition.h>
@ -21,10 +22,10 @@ private:
using DataPartPtr = std::shared_ptr<const DataPart>; using DataPartPtr = std::shared_ptr<const DataPart>;
public: public:
PartitionPruner(const KeyDescription & partition_key_, const SelectQueryInfo & query_info, const Context & context, bool strict) PartitionPruner(const KeyDescription & partition_key_, const ColumnsDescription & columns, const SelectQueryInfo & query_info, const Context & context, bool strict)
: partition_key(partition_key_) : partition_key(partition_key_)
, partition_condition( , partition_condition(
query_info, context, partition_key.column_names, partition_key.expression, true /* single_point */, strict) query_info, context, columns, partition_key.column_names, partition_key.expression, true /* single_point */, strict)
, useless(partition_condition.alwaysUnknownOrTrue()) , useless(partition_condition.alwaysUnknownOrTrue())
{ {
} }

View File

@ -210,7 +210,7 @@ std::optional<UInt64> StorageMergeTree::totalRows() const
std::optional<UInt64> StorageMergeTree::totalRowsByPartitionPredicate(const SelectQueryInfo & query_info, const Context & context) const std::optional<UInt64> StorageMergeTree::totalRowsByPartitionPredicate(const SelectQueryInfo & query_info, const Context & context) const
{ {
auto metadata_snapshot = getInMemoryMetadataPtr(); auto metadata_snapshot = getInMemoryMetadataPtr();
PartitionPruner partition_pruner(metadata_snapshot->getPartitionKey(), query_info, context, true /* strict */); PartitionPruner partition_pruner(metadata_snapshot->getPartitionKey(), metadata_snapshot->getColumns(), query_info, context, true /* strict */);
if (partition_pruner.isUseless()) if (partition_pruner.isUseless())
return {}; return {};
size_t res = 0; size_t res = 0;

View File

@ -3730,7 +3730,7 @@ std::optional<UInt64> StorageReplicatedMergeTree::totalRows() const
std::optional<UInt64> StorageReplicatedMergeTree::totalRowsByPartitionPredicate(const SelectQueryInfo & query_info, const Context & context) const std::optional<UInt64> StorageReplicatedMergeTree::totalRowsByPartitionPredicate(const SelectQueryInfo & query_info, const Context & context) const
{ {
auto metadata_snapshot = getInMemoryMetadataPtr(); auto metadata_snapshot = getInMemoryMetadataPtr();
PartitionPruner partition_pruner(metadata_snapshot->getPartitionKey(), query_info, context, true /* strict */); PartitionPruner partition_pruner(metadata_snapshot->getPartitionKey(), metadata_snapshot->getColumns(), query_info, context, true /* strict */);
if (partition_pruner.isUseless()) if (partition_pruner.isUseless())
return {}; return {};
size_t res = 0; size_t res = 0;

View File

@ -0,0 +1,17 @@
test-partition-prune
1
1
1
1
1
test-join
1
1
alias2alias
1
1
1
1
1
1
1

View File

@ -0,0 +1,70 @@
DROP TABLE IF EXISTS table_with_alias_column;
CREATE TABLE table_with_alias_column
(
`timestamp` DateTime,
`value` UInt64,
`day` Date ALIAS toDate(timestamp),
`day1` Date ALIAS day + 1,
`day2` Date ALIAS day1 + 1,
`time` DateTime ALIAS timestamp
)
ENGINE = MergeTree
PARTITION BY toYYYYMMDD(timestamp)
ORDER BY timestamp;
INSERT INTO table_with_alias_column(timestamp, value) SELECT toDateTime('2020-01-01 12:00:00'), 1 FROM numbers(1000);
INSERT INTO table_with_alias_column(timestamp, value) SELECT toDateTime('2020-01-02 12:00:00'), 1 FROM numbers(1000);
INSERT INTO table_with_alias_column(timestamp, value) SELECT toDateTime('2020-01-03 12:00:00'), 1 FROM numbers(1000);
SELECT 'test-partition-prune';
SELECT COUNT() = 1000 FROM table_with_alias_column WHERE day = '2020-01-01' SETTINGS max_rows_to_read = 1000;
SELECT t = '2020-01-03' FROM (SELECT day as t FROM table_with_alias_column WHERE t = '2020-01-03' GROUP BY t SETTINGS max_rows_to_read = 1000);
SELECT COUNT() = 1000 FROM table_with_alias_column WHERE day = '2020-01-01' UNION ALL select 1 from numbers(1) SETTINGS max_rows_to_read = 1001;
SELECT COUNT() = 0 FROM (SELECT toDate('2019-01-01') as day, day as t FROM table_with_alias_column PREWHERE t = '2020-01-03' WHERE t = '2020-01-03' GROUP BY t );
SELECT 'test-join';
SELECT day = '2020-01-03'
FROM
(
SELECT toDate('2020-01-03') AS day
FROM numbers(1)
) AS a
INNER JOIN
(
SELECT day
FROM table_with_alias_column
WHERE day = '2020-01-03'
GROUP BY day SETTINGS max_rows_to_read = 1000
) AS b ON a.day = b.day;
SELECT day = '2020-01-01'
FROM
(
SELECT day
FROM table_with_alias_column
WHERE day = '2020-01-01'
GROUP BY day SETTINGS max_rows_to_read = 1001
) AS a
INNER JOIN
(
SELECT toDate('2020-01-01') AS day
FROM numbers(1)
) AS b ON a.day = b.day;
SELECT 'alias2alias';
SELECT COUNT() = 1000 FROM table_with_alias_column WHERE day1 = '2020-01-02' SETTINGS max_rows_to_read = 1000;
SELECT t = '2020-01-03' FROM (SELECT day1 as t FROM table_with_alias_column WHERE t = '2020-01-03' GROUP BY t SETTINGS max_rows_to_read = 1000);
SELECT t = '2020-01-03' FROM (SELECT day2 as t FROM table_with_alias_column WHERE t = '2020-01-03' GROUP BY t SETTINGS max_rows_to_read = 1000);
SELECT COUNT() = 1000 FROM table_with_alias_column WHERE day1 = '2020-01-03' UNION ALL select 1 from numbers(1) SETTINGS max_rows_to_read = 1001;
SELECT COUNT() = 0 FROM (SELECT toDate('2019-01-01') as day1, day1 as t FROM table_with_alias_column PREWHERE t = '2020-01-03' WHERE t = '2020-01-03' GROUP BY t );
SELECT day1 = '2020-01-04' FROM table_with_alias_column PREWHERE day1 = '2020-01-04' WHERE day1 = '2020-01-04' GROUP BY day1 SETTINGS max_rows_to_read = 1000;
DROP TABLE table_with_alias_column;