Merge pull request #71866 from ClickHouse/fix_index_load_for_limit_1

Do not load indexes if key condition in not useful

src/Storages/MergeTree/MergeTreeDataSelectExecutor.cpp

@@ -1045,7 +1045,6 @@ MarkRanges MergeTreeDataSelectExecutor::markRangesFromPKRange(
     MarkRanges res;
 
     size_t marks_count = part->index_granularity.getMarksCount();
-    const auto & index = part->getIndex();
     if (marks_count == 0)
         return res;
 
@@ -1073,14 +1072,19 @@ MarkRanges MergeTreeDataSelectExecutor::markRangesFromPKRange(
     auto index_columns = std::make_shared<ColumnsWithTypeAndName>();
     const auto & key_indices = key_condition.getKeyIndices();
     DataTypes key_types;
-    for (size_t i : key_indices)
+    if (!key_indices.empty())
     {
-        if (i < index->size())
-            index_columns->emplace_back(index->at(i), primary_key.data_types[i], primary_key.column_names[i]);
-        else
-            index_columns->emplace_back(); /// The column of the primary key was not loaded in memory - we'll skip it.
+        const auto & index = part->getIndex();
 
-        key_types.emplace_back(primary_key.data_types[i]);
+        for (size_t i : key_indices)
+        {
+            if (i < index->size())
+                index_columns->emplace_back(index->at(i), primary_key.data_types[i], primary_key.column_names[i]);
+            else
+                index_columns->emplace_back(); /// The column of the primary key was not loaded in memory - we'll skip it.
+
+            key_types.emplace_back(primary_key.data_types[i]);
+        }
     }
 
     /// If there are no monotonic functions, there is no need to save block reference.

tests/queries/0_stateless/02993_lazy_index_loading.reference

@@ -1,4 +1,18 @@
-100000000	100000000
-0	0
+Row 1:
+──────
+round(primary_key_bytes_in_memory, -7):           100000000 -- 100.00 million
+round(primary_key_bytes_in_memory_allocated, -7): 100000000 -- 100.00 million
+Row 1:
+──────
+primary_key_bytes_in_memory:           0
+primary_key_bytes_in_memory_allocated: 0
 1
-100000000	100000000
+Row 1:
+──────
+primary_key_bytes_in_memory:           0
+primary_key_bytes_in_memory_allocated: 0
+1
+Row 1:
+──────
+round(primary_key_bytes_in_memory, -7):           100000000 -- 100.00 million
+round(primary_key_bytes_in_memory_allocated, -7): 100000000 -- 100.00 million

tests/queries/0_stateless/02993_lazy_index_loading.sql

@@ -3,17 +3,26 @@ CREATE TABLE test (s String) ENGINE = MergeTree ORDER BY s SETTINGS index_granul
 
 SET optimize_trivial_insert_select = 1;
 INSERT INTO test SELECT randomString(1000) FROM numbers(100000);
-SELECT round(primary_key_bytes_in_memory, -7), round(primary_key_bytes_in_memory_allocated, -7) FROM system.parts WHERE database = currentDatabase() AND table = 'test';
+SELECT round(primary_key_bytes_in_memory, -7), round(primary_key_bytes_in_memory_allocated, -7) FROM system.parts WHERE database = currentDatabase() AND table = 'test' FORMAT Vertical;
 
 DETACH TABLE test;
 SET max_memory_usage = '50M';
 ATTACH TABLE test;
 
-SELECT primary_key_bytes_in_memory, primary_key_bytes_in_memory_allocated FROM system.parts WHERE database = currentDatabase() AND table = 'test';
+SELECT primary_key_bytes_in_memory, primary_key_bytes_in_memory_allocated FROM system.parts WHERE database = currentDatabase() AND table = 'test' FORMAT Vertical;
 
 SET max_memory_usage = '200M';
+
+-- Run a query that doesn use indexes
 SELECT s != '' FROM test LIMIT 1;
 
-SELECT round(primary_key_bytes_in_memory, -7), round(primary_key_bytes_in_memory_allocated, -7) FROM system.parts WHERE database = currentDatabase() AND table = 'test';
+-- Check that index was not loaded
+SELECT primary_key_bytes_in_memory, primary_key_bytes_in_memory_allocated FROM system.parts WHERE database = currentDatabase() AND table = 'test' FORMAT Vertical;
+
+-- Run a query that uses PK index
+SELECT s != '' FROM test WHERE s < '9999999999' LIMIT 1;
+
+-- Check that index was loaded
+SELECT round(primary_key_bytes_in_memory, -7), round(primary_key_bytes_in_memory_allocated, -7) FROM system.parts WHERE database = currentDatabase() AND table = 'test' FORMAT Vertical;
 
 DROP TABLE test;

tests/queries/0_stateless/03127_system_unload_primary_key_table.reference

@@ -5,9 +5,19 @@
 100000000	100000000
 0	0
 0	0
+Query that does not use index for table `test`
+1
+0	0
+0	0
+Query that uses index in for table `test`
 1
 100000000	100000000
 0	0
+Query that does not use index for table `test2`
+1
+100000000	100000000
+0	0
+Query that uses index for table `test2`
 1
 100000000	100000000
 100000000	100000000

tests/queries/0_stateless/03127_system_unload_primary_key_table.sql

@@ -16,8 +16,18 @@ SELECT round(primary_key_bytes_in_memory, -7), round(primary_key_bytes_in_memory
 SYSTEM UNLOAD PRIMARY KEY {CLICKHOUSE_DATABASE:Identifier}.test2;
 SELECT round(primary_key_bytes_in_memory, -7), round(primary_key_bytes_in_memory_allocated, -7) FROM system.parts WHERE database = currentDatabase() AND table IN ('test', 'test2');
 
+SELECT 'Query that does not use index for table `test`';
 SELECT s != '' FROM test LIMIT 1;
 SELECT round(primary_key_bytes_in_memory, -7), round(primary_key_bytes_in_memory_allocated, -7) FROM system.parts WHERE database = currentDatabase() AND table IN ('test', 'test2');
 
+SELECT 'Query that uses index in for table `test`';
+SELECT s != '' FROM test WHERE s < '99999999' LIMIT 1;
+SELECT round(primary_key_bytes_in_memory, -7), round(primary_key_bytes_in_memory_allocated, -7) FROM system.parts WHERE database = currentDatabase() AND table IN ('test', 'test2');
+
+SELECT 'Query that does not use index for table `test2`';
 SELECT s != '' FROM test2 LIMIT 1;
 SELECT round(primary_key_bytes_in_memory, -7), round(primary_key_bytes_in_memory_allocated, -7) FROM system.parts WHERE database = currentDatabase() AND table IN ('test', 'test2');
+
+SELECT 'Query that uses index for table `test2`';
+SELECT s != '' FROM test2 WHERE s < '99999999' LIMIT 1;
+SELECT round(primary_key_bytes_in_memory, -7), round(primary_key_bytes_in_memory_allocated, -7) FROM system.parts WHERE database = currentDatabase() AND table IN ('test', 'test2');