From 159797ab236d8a7bfb4c8fc5b2a3bcf258dd89f0 Mon Sep 17 00:00:00 2001
From: ivanmorozov333 <ivanmorozov@hotmail.com>
Date: Tue, 16 Jan 2024 11:21:54 +0300
Subject: [PATCH 1/2] fix case removed column for stats usage

---
 .../tx/columnshard/engines/changes/general_compaction.cpp | 2 --
 ydb/core/tx/columnshard/engines/portions/common.h         | 8 ++++++++
 ydb/core/tx/columnshard/engines/portions/portion_info.h   | 4 +++-
 3 files changed, 11 insertions(+), 3 deletions(-)

diff --git a/ydb/core/tx/columnshard/engines/changes/general_compaction.cpp b/ydb/core/tx/columnshard/engines/changes/general_compaction.cpp
index 08e4fad95c66..8c095bbe091b 100644
--- a/ydb/core/tx/columnshard/engines/changes/general_compaction.cpp
+++ b/ydb/core/tx/columnshard/engines/changes/general_compaction.cpp
@@ -18,7 +18,6 @@ namespace NKikimr::NOlap::NCompaction {
 
 void TGeneralCompactColumnEngineChanges::BuildAppendedPortionsByFullBatches(TConstructionContext& context) noexcept {
     std::vector<TPortionInfoWithBlobs> portions = TPortionInfoWithBlobs::RestorePortions(SwitchedPortions, Blobs);
-    Blobs.clear();
     std::vector<std::shared_ptr<arrow::RecordBatch>> batchResults;
     auto resultSchema = context.SchemaVersions.GetLastSchema();
     {
@@ -45,7 +44,6 @@ void TGeneralCompactColumnEngineChanges::BuildAppendedPortionsByFullBatches(TCon
 
 void TGeneralCompactColumnEngineChanges::BuildAppendedPortionsByChunks(TConstructionContext& context) noexcept {
     std::vector<TPortionInfoWithBlobs> portions = TPortionInfoWithBlobs::RestorePortions(SwitchedPortions, Blobs);
-    Blobs.clear();
     static const TString portionIdFieldName = "$$__portion_id";
     static const TString portionRecordIndexFieldName = "$$__portion_record_idx";
     static const std::shared_ptr<arrow::Field> portionIdField = std::make_shared<arrow::Field>(portionIdFieldName, std::make_shared<arrow::UInt16Type>());
diff --git a/ydb/core/tx/columnshard/engines/portions/common.h b/ydb/core/tx/columnshard/engines/portions/common.h
index b0305fa53e9a..311563918469 100644
--- a/ydb/core/tx/columnshard/engines/portions/common.h
+++ b/ydb/core/tx/columnshard/engines/portions/common.h
@@ -8,6 +8,14 @@ class TChunkAddress {
     YDB_READONLY(ui32, ColumnId, 0);
     YDB_READONLY(ui16, Chunk, 0);
 public:
+    ui32 GetEntityId() const {
+        return ColumnId;
+    }
+
+    ui32 GetChunkIdx() const {
+        return Chunk;
+    }
+
     TChunkAddress(const ui32 columnId, const ui16 chunk)
         : ColumnId(columnId)
         , Chunk(chunk) {
diff --git a/ydb/core/tx/columnshard/engines/portions/portion_info.h b/ydb/core/tx/columnshard/engines/portions/portion_info.h
index 19ba5e3c1df7..00e4fd205aef 100644
--- a/ydb/core/tx/columnshard/engines/portions/portion_info.h
+++ b/ydb/core/tx/columnshard/engines/portions/portion_info.h
@@ -97,7 +97,9 @@ class TPortionInfo {
     TSerializationStats GetSerializationStat(const ISnapshotSchema& schema) const {
         TSerializationStats result;
         for (auto&& i : Records) {
-            result.AddStat(i.GetSerializationStat(schema.GetFieldByColumnIdVerified(i.ColumnId)->name()));
+            if (schema.GetFieldByColumnId(i.ColumnId)) {
+                result.AddStat(i.GetSerializationStat(schema.GetFieldByColumnIdVerified(i.ColumnId)->name()));
+            }
         }
         return result;
     }

From 7c8565aada8ccbeb15ab7c3729c680dd82d171fe Mon Sep 17 00:00:00 2001
From: ivanmorozov333 <ivanmorozov@hotmail.com>
Date: Tue, 16 Jan 2024 11:27:05 +0300
Subject: [PATCH 2/2] additional logging

---
 ydb/core/tx/columnshard/engines/scheme/abstract_scheme.cpp | 6 ++++++
 ydb/core/tx/columnshard/engines/scheme/abstract_scheme.h   | 6 +-----
 2 files changed, 7 insertions(+), 5 deletions(-)

diff --git a/ydb/core/tx/columnshard/engines/scheme/abstract_scheme.cpp b/ydb/core/tx/columnshard/engines/scheme/abstract_scheme.cpp
index f138fc6b5f3f..13dc7a35f1e9 100644
--- a/ydb/core/tx/columnshard/engines/scheme/abstract_scheme.cpp
+++ b/ydb/core/tx/columnshard/engines/scheme/abstract_scheme.cpp
@@ -110,4 +110,10 @@ ui32 ISnapshotSchema::GetColumnId(const std::string& columnName) const {
     return *id;
 }
 
+std::shared_ptr<arrow::Field> ISnapshotSchema::GetFieldByColumnIdVerified(const ui32 columnId) const {
+    auto result = GetFieldByColumnId(columnId);
+    AFL_VERIFY(result)("event", "unknown_column")("column_id", columnId);
+    return result;
+}
+
 }
diff --git a/ydb/core/tx/columnshard/engines/scheme/abstract_scheme.h b/ydb/core/tx/columnshard/engines/scheme/abstract_scheme.h
index 1ff127bee20a..9232edcf1dfe 100644
--- a/ydb/core/tx/columnshard/engines/scheme/abstract_scheme.h
+++ b/ydb/core/tx/columnshard/engines/scheme/abstract_scheme.h
@@ -52,11 +52,7 @@ class ISnapshotSchema {
     ui32 GetColumnId(const std::string& columnName) const;
     std::shared_ptr<arrow::Field> GetFieldByIndex(const int index) const;
     std::shared_ptr<arrow::Field> GetFieldByColumnId(const ui32 columnId) const;
-    std::shared_ptr<arrow::Field> GetFieldByColumnIdVerified(const ui32 columnId) const {
-        auto result = GetFieldByColumnId(columnId);
-        Y_ABORT_UNLESS(result);
-        return result;
-    }
+    std::shared_ptr<arrow::Field> GetFieldByColumnIdVerified(const ui32 columnId) const;
 
     TString DebugString() const {
         return DoDebugString();