Merge pull request ClickHouse#80597 from zoomxi/keycondition_improve

Enable binary search for wrapped primary keys with always monotonic function chains

Author: pamarcos

src/Storages/MergeTree/KeyCondition.cpp

@@ -2892,9 +2892,41 @@ std::optional<Range> KeyCondition::applyMonotonicFunctionsChainToRange(
 // This allows to use a more efficient lookup with no extra reads.
 bool KeyCondition::matchesExactContinuousRange() const
 {
-    // Not implemented yet.
-    if (hasMonotonicFunctionsChain())
-        return false;
+    const Field field{};
+    auto is_always_monotonic_chain = [&field](const std::vector<FunctionBasePtr> & chain)
+    {
+        for (const auto & func : chain)
+        {
+            if (!func || !func->hasInformationAboutMonotonicity())
+                return false;
+
+            const auto & types = func->getArgumentTypes();
+            if (types.empty() || !types.front())
+                return false;
+
+            const auto monotonicity = func->getMonotonicityForRange(*types.front(), field, field);
+            if (!monotonicity.is_always_monotonic)
+                return false;
+        }
+
+        return true;
+    };
+
+    for (const auto & elem : rpn)
+    {
+        if (!elem.monotonic_functions_chain.empty() && !is_always_monotonic_chain(elem.monotonic_functions_chain))
+            return false;
+
+        if (elem.set_index)
+        {
+            if (elem.function != RPNElement::Function::FUNCTION_IN_SET || elem.set_index->size() != 1)
+                return false;
+
+            for (const auto & mapping : elem.set_index->getIndexesMapping())
+                if (!mapping.functions.empty() && !is_always_monotonic_chain(mapping.functions))
+                    return false;
+        }
+    }
 
     enum Constraint
     {
@@ -2936,6 +2968,11 @@ bool KeyCondition::matchesExactContinuousRange() const
             continue;
         }
 
+        if (element.function == RPNElement::Function::ALWAYS_TRUE)
+        {
+            continue;
+        }
+
         return false;
     }
 

tests/queries/0_stateless/03518_key_condition_binary_search.reference

@@ -0,0 +1,2 @@
+3
+1

tests/queries/0_stateless/03518_key_condition_binary_search.sh

@@ -0,0 +1,31 @@
+#!/usr/bin/env bash
+
+CUR_DIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
+# shellcheck source=../shell_config.sh
+. "$CUR_DIR"/../shell_config.sh
+
+$CLICKHOUSE_CLIENT -n -q "
+DROP TABLE IF EXISTS t;
+CREATE TABLE t
+(
+    c Enum8('Zero' = 0, 'One' = 1, 'Two' = 2, 'Three' = 3, 'Four' = 4, 'Five' = 5)
+)
+ENGINE = MergeTree
+ORDER BY c;
+INSERT INTO t values('One');
+SELECT * FROM t WHERE c = 1 settings send_logs_level='trace';
+SELECT * FROM t WHERE c = 'One' settings send_logs_level='trace';
+SELECT * FROM t WHERE c = 1 and 1 = 1 settings send_logs_level='trace';
+" 2>&1 | grep -c "binary search"
+
+$CLICKHOUSE_CLIENT -n -q "
+DROP TABLE IF EXISTS t1;
+CREATE TABLE t1
+(
+    timestamp DateTime64(3, 'Asia/Shanghai')
+)
+ENGINE = MergeTree
+ORDER BY timestamp;
+INSERT INTO t1 VALUES ('2025-05-21 00:00:00');
+SELECT * FROM t1 WHERE toDayOfMonth(timestamp) = 1 settings send_logs_level='trace';
+" 2>&1 | grep -c "generic exclusion search"