YQL-19616: Fix TRegexLexer performance

Fix `TRegexLexer` performance. Now it is just 2 times slower than a reference ANTLR implementation on Release mode, so merged regexes are 3 times better than scan&compare.

![image](https://github.com/user-attachments/assets/4e0cb27a-491d-4dbd-b10a-5725ffa6d902)

---

- Related to `YQL-19616`
- Related to https://github.com/ydb-platform/ydb/issues/15129
- Related to https://github.com/vityaman/ydb/issues/42

---

Pull Request resolved: https://github.com/ytsaurus/ytsaurus/pull/1278
commit_hash:1529f641172fea13f0d33fbfd06a4827c6efde01

1 files changed, 11 insertions, 2 deletions

diff --git a/yql/essentials/sql/v1/reflect/sql_reflect.cpp b/yql/essentials/sql/v1/reflect/sql_reflect.cpp
index 262209cfc39..5e652db4fc0 100644
--- a/yql/essentials/sql/v1/reflect/sql_reflect.cpp
+++ b/yql/essentials/sql/v1/reflect/sql_reflect.cpp
@@ -1,9 +1,11 @@
 #include "sql_reflect.h"
 
 #include <library/cpp/resource/resource.h>
+#include <library/cpp/case_insensitive_string/case_insensitive_string.h>
 
 #include <util/string/split.h>
 #include <util/string/strip.h>
+#include <util/charset/utf8.h>
 
 namespace NSQLReflect {
 
@@ -15,13 +17,20 @@ namespace NSQLReflect {
     const TStringBuf SectionOther = "//! section:other";
     const TStringBuf FragmentPrefix = "fragment ";
 
-    const TStringBuf TLexerGrammar::KeywordBlock(const TStringBuf name) {
+    const TStringBuf TLexerGrammar::KeywordBlockByName(const TStringBuf name Y_LIFETIME_BOUND) {
         if (name == "TSKIP") {
             return "SKIP";
         }
         return name;
     }
 
+    const TString TLexerGrammar::KeywordNameByBlock(const TStringBuf block) {
+        if (TCaseInsensitiveStringBuf(block) == "SKIP") {
+            return "TSKIP";
+        }
+        return ToUpperUTF8(block);
+    }
+
     TVector<TString> GetResourceLines(const TStringBuf key) {
         TString text;
         Y_ENSURE(NResource::FindExact(key, &text));
@@ -133,7 +142,7 @@ namespace NSQLReflect {
         SubstGlobal(block, "'", "");
         SubstGlobal(block, " ", "");
 
-        Y_ENSURE(name == block || (name == "TSKIP" && block == TLexerGrammar::KeywordBlock("TSKIP")));
+        Y_ENSURE(name == block || (name == "TSKIP" && block == TLexerGrammar::KeywordBlockByName("TSKIP")));
         grammar.KeywordNames.emplace(std::move(name));
     }