KIKIMR-25621: Introduce multi-column statistics

#### Introduce multi-column statistics support ✎

- Adds support for defining multi-column statistics in table schemas, allowing users to specify statistics for combinations of columns rather than individual columns
- Introduces new SQL syntax for CREATE TABLE and ALTER TABLE statements to declare statistics with optional type specifications
- Implements validation for statistics definitions, ensuring column references are valid and names are unique
- Extends the query processing pipeline to handle statistics declarations during table creation and modification operations
- Updates syntax highlighting and parsing rules to recognize the new statistics keyword and related constructs
- Adds comprehensive unit tests covering various statistics scenarios including multi-column definitions, type specifications, and validation rules

<a href="https://nda.ya.ru/t/qa0kX64r7DqvtN"><font size="2">Autodescription by Yandex Code Assistant</font></a>
commit_hash:5ca3f8721d3e3660f2bb890c1408e420cf7c3e81

1 files changed, 30 insertions, 0 deletions

diff --git a/yql/essentials/sql/v1/sql_translation.cpp b/yql/essentials/sql/v1/sql_translation.cpp
index b95e1c61b56..17596ab4d9c 100644
--- a/yql/essentials/sql/v1/sql_translation.cpp
+++ b/yql/essentials/sql/v1/sql_translation.cpp
@@ -790,6 +790,28 @@ bool TSqlTranslation::CreateTableIndex(const TRule_table_index& node, TVector<TI
     return true;
 }
 
+bool TSqlTranslation::CreateTableStatistics(const TRule_table_statistics& node, TVector<TStatisticsDescription>& statistics) {
+    statistics.emplace_back(IdEx(node.GetRule_an_id2(), *this));
+    auto& stat = statistics.back();
+
+    // ON (columns)
+    stat.Columns.emplace_back(IdEx(node.GetRule_an_id_schema5(), *this));
+    for (const auto& block : node.GetBlock6()) {
+        stat.Columns.emplace_back(IdEx(block.GetRule_an_id_schema2(), *this));
+    }
+
+    // WITH (types) — optional; omission means "all supported statistic types"
+    if (node.HasBlock8()) {
+        const auto& withTypes = node.GetBlock8().GetRule_with_statistics_types1();
+        stat.Types.emplace_back(IdEx(withTypes.GetRule_an_id_or_type3(), *this));
+        for (const auto& block : withTypes.GetBlock4()) {
+            stat.Types.emplace_back(IdEx(block.GetRule_an_id_or_type2(), *this));
+        }
+    }
+
+    return true;
+}
+
 bool TSqlTranslation::ParseDatabaseSettings(const TRule_database_settings& in, THashMap<TString, TNodePtr>& out) {
     if (!ParseDatabaseSetting(in.GetRule_database_setting1(), out)) {
         return false;
@@ -2156,6 +2178,14 @@ bool TSqlTranslation::CreateTableEntry(const TRule_create_table_entry& node, TCr
             params.Columns.push_back({.Pos = pos, .Name = name, .Nullable = true});
             break;
         }
+        case TRule_create_table_entry::kAltCreateTableEntry7: {
+            // table_statistics
+            auto& table_statistics = node.GetAlt_create_table_entry7().GetRule_table_statistics1();
+            if (!CreateTableStatistics(table_statistics, params.Statistics)) {
+                return false;
+            }
+            break;
+        }
         case NSQLv1Generated::TRule_create_table_entry::ALT_NOT_SET:
             YQL_ENSURE(false, "Unreachable");
     }