From 456c37b57c8c7c3559a2a4e51f3b43335651a1cb Mon Sep 17 00:00:00 2001
From: ivanmorozov333 <ivanmorozov@hotmail.com>
Date: Tue, 17 Sep 2024 11:03:46 +0300
Subject: [PATCH] fix addressing and additional logging

---
 .../engines/changes/compaction/sparsed/logic.h           | 6 ++++--
 ydb/library/formats/arrow/accessor/abstract/accessor.cpp | 2 +-
 ydb/library/formats/arrow/accessor/abstract/accessor.h   | 9 +++++++--
 3 files changed, 12 insertions(+), 5 deletions(-)

diff --git a/ydb/core/tx/columnshard/engines/changes/compaction/sparsed/logic.h b/ydb/core/tx/columnshard/engines/changes/compaction/sparsed/logic.h
index 35a3649d61f3..f9a96d5bf974 100644
--- a/ydb/core/tx/columnshard/engines/changes/compaction/sparsed/logic.h
+++ b/ydb/core/tx/columnshard/engines/changes/compaction/sparsed/logic.h
@@ -58,7 +58,9 @@ class TSparsedMerger: public IColumnMerger {
 
         void InitArrays(const ui32 position) {
             AFL_VERIFY(!ChunkAddress || ChunkFinishPosition <= position);
-            ChunkAddress = CurrentChunkedArray->GetChunk(ChunkAddress, position);
+            AFL_VERIFY(CurrentOwnedArray->GetAddress().GetGlobalStartPosition() <= position)("pos", position)(
+                "global", CurrentOwnedArray->GetAddress().GetGlobalStartPosition());
+            ChunkAddress = CurrentChunkedArray->GetChunk(ChunkAddress, position - CurrentOwnedArray->GetAddress().GetGlobalStartPosition());
             AFL_VERIFY(ChunkAddress);
             ChunkStartPosition = CurrentOwnedArray->GetAddress().GetGlobalStartPosition() + ChunkAddress->GetAddress().GetGlobalStartPosition();
             ChunkFinishPosition =
@@ -83,7 +85,7 @@ class TSparsedMerger: public IColumnMerger {
                     return {};
                 }
                 if (ChunkFinishPosition <= currentIndex) {
-                    InitArrays(currentGlobalPosition);
+                    InitArrays(currentIndex);
                     continue;
                 }
                 for (; currentIndex < ChunkFinishPosition; ++currentIndex) {
diff --git a/ydb/library/formats/arrow/accessor/abstract/accessor.cpp b/ydb/library/formats/arrow/accessor/abstract/accessor.cpp
index 9d4c33efa42e..c58ed0e28373 100644
--- a/ydb/library/formats/arrow/accessor/abstract/accessor.cpp
+++ b/ydb/library/formats/arrow/accessor/abstract/accessor.cpp
@@ -45,7 +45,7 @@ std::shared_ptr<arrow::ChunkedArray> IChunkedArray::Slice(const ui32 offset, con
 
 NKikimr::NArrow::NAccessor::IChunkedArray::TFullDataAddress IChunkedArray::GetChunk(
     const std::optional<TAddressChain>& chunkCurrent, const ui64 position) const {
-    AFL_VERIFY(position < GetRecordsCount());
+    AFL_VERIFY(position < GetRecordsCount())("pos", position)("records", GetRecordsCount())("current", chunkCurrent ? chunkCurrent->DebugString() : Default<TString>());
     std::optional<TCommonChunkAddress> address;
 
     if (IsDataOwner()) {
diff --git a/ydb/library/formats/arrow/accessor/abstract/accessor.h b/ydb/library/formats/arrow/accessor/abstract/accessor.h
index c8e7979618dd..934ca2fddea4 100644
--- a/ydb/library/formats/arrow/accessor/abstract/accessor.h
+++ b/ydb/library/formats/arrow/accessor/abstract/accessor.h
@@ -104,8 +104,13 @@ class IChunkedArray {
         }
 
         TString DebugString() const {
-            return TStringBuilder() << "start=" << GlobalStartPosition << ";finish=" << GlobalFinishPosition
-                                    << ";addresses_count=" << Addresses.size() << ";";
+            TStringBuilder sb;
+            sb << "start=" << GlobalStartPosition << ";finish=" << GlobalFinishPosition
+                << ";addresses_count=" << Addresses.size() << ";";
+            for (auto&& i : Addresses) {
+                sb << "addresses=" << i.DebugString() << ";";
+            }
+            return sb;
         }
     };