stash

src/Interpreters/ActionsDAG.cpp

@@ -1527,15 +1527,12 @@ ActionsDAG::SplitResult ActionsDAG::splitActionsBeforeArrayJoin(const NameSet &
     return res;
 }
 
-ActionsDAG::SplitResult ActionsDAG::splitActionsBySortingDescription(const SortDescription & sort_description) const
+ActionsDAG::SplitResult ActionsDAG::splitActionsBySortingDescription(const NameSet & sort_columns) const
 {
     std::unordered_set<const Node *> split_nodes;
-    for (const auto & sort_column : sort_description)
-    {
-        const auto * node = tryFindInIndex(sort_column.column_name);
-        if (node)
+    for (const auto & sort_column : sort_columns)
+        if (const auto * node = tryFindInIndex(sort_column))
             split_nodes.insert(node);
-    }
     auto res = split(split_nodes);
     res.second->project_input = project_input;
     return res;

src/Interpreters/ActionsDAG.h

@@ -7,8 +7,6 @@
 
 #include "config_core.h"
 
-#include <Core/SortDescription.h>
-
 namespace DB
 {
 
@@ -276,8 +274,9 @@ public:
     /// Index of initial actions must contain column_name.
     SplitResult splitActionsForFilter(const std::string & column_name) const;
 
-    ///
-    SplitResult splitActionsBySortingDescription(const SortDescription & sort_description) const;
+    /// Splits actions into two parts. The first part contains all the calculations required to calculate sort_columns.
+    /// The second contains the rest.
+    SplitResult splitActionsBySortingDescription(const NameSet & sort_columns) const;
 
     /// Create actions which may calculate part of filter using only available_inputs.
     /// If nothing may be calculated, returns nullptr.

src/Processors/QueryPlan/Optimizations/Optimizations.h

@@ -44,7 +44,8 @@ size_t tryMergeExpressions(QueryPlan::Node * parent_node, QueryPlan::Nodes &);
 /// May split FilterStep and push down only part of it.
 size_t tryPushDownFilter(QueryPlan::Node * parent_node, QueryPlan::Nodes & nodes);
 
-///
+/// Move ExpressionStep up if possible.
+/// May split ExpressionStep and lift up only part of it.
 size_t tryExecuteFunctionsAfterSorting(QueryPlan::Node * parent_node, QueryPlan::Nodes & nodes);
 
 inline const auto & getOptimizations()

src/Processors/QueryPlan/Optimizations/liftUpFunctions.cpp

@@ -42,11 +42,14 @@ size_t tryExecuteFunctionsAfterSorting(QueryPlan::Node * parent_node, QueryPlan:
     if (!sorting_step || !expression_step)
         return 0;
 
-    const auto & sort_columns = sorting_step->getSortDescription();
+    NameSet sort_columns;
+    for (const auto & col : sorting_step->getSortDescription())
+        sort_columns.insert(col.column_name);
+
     const auto & expression = expression_step->getExpression();
 
     for (auto sc : sort_columns)
-        LOG_TRACE(&Poco::Logger::get("Optimizer"), "sort_columns: {}", sc.column_name);
+        LOG_TRACE(&Poco::Logger::get("Optimizer"), "sort_columns: {}", fmt::join(sort_columns, ", "));
 
     auto split_actions = expression->splitActionsBySortingDescription(sort_columns);
     LOG_TRACE(&Poco::Logger::get("Optimizer"), "source: {}", expression->dumpDAG());

src/Processors/QueryPlan/SortingStep.cpp

@@ -89,17 +89,15 @@ SortingStep::SortingStep(
     output_stream->sort_mode = DataStream::SortMode::Stream;
 }
 
-void SortingStep::updateInputStream(const DataStream & input_stream)
+void SortingStep::updateInputStream(DataStream input_stream)
 {
     input_streams.clear();
-    input_streams.emplace_back(input_stream);
+    input_streams.push_back(std::move(input_stream));
 }
 
 void SortingStep::updateOutputStream(Block result_header)
 {
-    if (input_streams.size() != 1)
-        throw std::runtime_error{"wasted"};
-    output_stream = createOutputStream(input_streams.at(0), result_header, getDataStreamTraits());
+    output_stream = createOutputStream(input_streams.front(), std::move(result_header), getDataStreamTraits());
 }
 
 void SortingStep::updateLimit(size_t limit_)

src/Processors/QueryPlan/SortingStep.h

@@ -49,7 +49,7 @@ public:
     /// Add limit or change it to lower value.
     void updateLimit(size_t limit_);
 
-    void updateInputStream(const DataStream & input_stream);
+    void updateInputStream(DataStream input_stream);
     void updateOutputStream(Block result_header);
 
     SortDescription getSortDescription() const { return result_description; }

tests/queries/0_stateless/01655_plan_optimizations.reference

@@ -148,7 +148,7 @@ Expression
 Limit
 Expression
 Sorting
-> Expression should be divided into two subnodes and only one of them could be moved after Sorting
+> Expression should be divided into two subexpressions and only one of them should be moved after Sorting
 Expression
 Limit
 Expression

tests/queries/0_stateless/01655_plan_optimizations.sh

@@ -202,7 +202,7 @@ echo "> the whole Expression node could be moved after Sorting"
 $CLICKHOUSE_CLIENT -q "
     explain select sipHash64(number) from numbers(100) order by number limit 5" |
     sed 's/ //g' | grep -o "^ *\(Expression\|Limit\|Sorting\)"
-echo "> Expression should be divided into two subnodes and only one of them could be moved after Sorting"
+echo "> Expression should be divided into two subexpressions and only one of them should be moved after Sorting"
 $CLICKHOUSE_CLIENT -q "
     explain select sipHash64(number) from numbers(100) order by number + 1 limit 5" |
     sed 's/ //g' | grep -o "^ *\(Expression\|Limit\|Sorting\)"