From 8fbacf358315530306ca43efc0b188bc78b4382f Mon Sep 17 00:00:00 2001
From: Bingyi Sun <sunbingyi1992@gmail.com>
Date: Fri, 14 Mar 2025 16:06:08 +0800
Subject: [PATCH] fix: Null expr does not work for json field (#40456)

issue: https://github.com/milvus-io/milvus/issues/40455

Signed-off-by: sunby <sunbingyi1992@gmail.com>
---
 internal/core/src/mmap/ChunkedColumn.h        | 20 ++++++++++++-------
 .../src/segcore/ChunkedSegmentSealedImpl.cpp  | 13 ++++++------
 internal/core/src/storage/ChunkCache.cpp      |  2 +-
 3 files changed, 21 insertions(+), 14 deletions(-)

diff --git a/internal/core/src/mmap/ChunkedColumn.h b/internal/core/src/mmap/ChunkedColumn.h
index 0136231b90..73fbdf2d8c 100644
--- a/internal/core/src/mmap/ChunkedColumn.h
+++ b/internal/core/src/mmap/ChunkedColumn.h
@@ -54,9 +54,7 @@ class ChunkedColumnBase : public ColumnBase {
     ChunkedColumnBase() = default;
     // memory mode ctor
     explicit ChunkedColumnBase(const FieldMeta& field_meta) {
-        if (field_meta.is_nullable()) {
-            nullable_ = true;
-        }
+        nullable_ = field_meta.is_nullable();
     }
 
     virtual ~ChunkedColumnBase() = default;
@@ -248,7 +246,9 @@ class ChunkedColumn : public ChunkedColumnBase {
         : ChunkedColumnBase(field_meta) {
     }
 
-    explicit ChunkedColumn(const std::vector<std::shared_ptr<Chunk>>& chunks) {
+    explicit ChunkedColumn(const FieldMeta& field_meta,
+                           const std::vector<std::shared_ptr<Chunk>>& chunks)
+        : ChunkedColumnBase(field_meta) {
         for (auto& chunk : chunks) {
             AddChunk(chunk);
         }
@@ -272,7 +272,9 @@ class ChunkedSparseFloatColumn : public ChunkedColumnBase {
     }
 
     explicit ChunkedSparseFloatColumn(
-        const std::vector<std::shared_ptr<Chunk>>& chunks) {
+        const FieldMeta& field_meta,
+        const std::vector<std::shared_ptr<Chunk>>& chunks)
+        : ChunkedColumnBase(field_meta) {
         for (auto& chunk : chunks) {
             AddChunk(chunk);
         }
@@ -317,7 +319,9 @@ class ChunkedVariableColumn : public ChunkedColumnBase {
     }
 
     explicit ChunkedVariableColumn(
-        const std::vector<std::shared_ptr<Chunk>>& chunks) {
+        const FieldMeta& field_meta,
+        const std::vector<std::shared_ptr<Chunk>>& chunks)
+        : ChunkedColumnBase(field_meta) {
         for (auto& chunk : chunks) {
             AddChunk(chunk);
         }
@@ -397,7 +401,9 @@ class ChunkedArrayColumn : public ChunkedColumnBase {
     }
 
     explicit ChunkedArrayColumn(
-        const std::vector<std::shared_ptr<Chunk>>& chunks) {
+        const FieldMeta& field_meta,
+        const std::vector<std::shared_ptr<Chunk>>& chunks)
+        : ChunkedColumnBase(field_meta) {
         for (auto& chunk : chunks) {
             AddChunk(chunk);
         }
diff --git a/internal/core/src/segcore/ChunkedSegmentSealedImpl.cpp b/internal/core/src/segcore/ChunkedSegmentSealedImpl.cpp
index 8becdddc7d..9a3fc55b7c 100644
--- a/internal/core/src/segcore/ChunkedSegmentSealedImpl.cpp
+++ b/internal/core/src/segcore/ChunkedSegmentSealedImpl.cpp
@@ -531,7 +531,7 @@ ChunkedSegmentSealedImpl::MapFieldData(const FieldId field_id,
                 //     DEFAULT_MMAP_VRCOL_BLOCK_SIZE);
                 auto var_column =
                     std::make_shared<ChunkedVariableColumn<std::string>>(
-                        chunks);
+                        field_meta, chunks);
                 // var_column->Seal(std::move(indices));
                 column = std::move(var_column);
                 break;
@@ -539,21 +539,22 @@ ChunkedSegmentSealedImpl::MapFieldData(const FieldId field_id,
             case milvus::DataType::JSON: {
                 auto var_column =
                     std::make_shared<ChunkedVariableColumn<milvus::Json>>(
-                        chunks);
+                        field_meta, chunks);
                 // var_column->Seal(std::move(indices));
                 column = std::move(var_column);
                 break;
             }
             case milvus::DataType::ARRAY: {
-                auto arr_column = std::make_shared<ChunkedArrayColumn>(chunks);
+                auto arr_column =
+                    std::make_shared<ChunkedArrayColumn>(field_meta, chunks);
                 // arr_column->Seal(std::move(indices),
                 //                  std::move(element_indices));
                 column = std::move(arr_column);
                 break;
             }
             case milvus::DataType::VECTOR_SPARSE_FLOAT: {
-                auto sparse_column =
-                    std::make_shared<ChunkedSparseFloatColumn>(chunks);
+                auto sparse_column = std::make_shared<ChunkedSparseFloatColumn>(
+                    field_meta, chunks);
                 // sparse_column->Seal(std::move(indices));
                 column = std::move(sparse_column);
                 break;
@@ -564,7 +565,7 @@ ChunkedSegmentSealedImpl::MapFieldData(const FieldId field_id,
             }
         }
     } else {
-        column = std::make_shared<ChunkedColumn>(chunks);
+        column = std::make_shared<ChunkedColumn>(field_meta, chunks);
     }
 
     // column->SetValidData(std::move(valid_data));
diff --git a/internal/core/src/storage/ChunkCache.cpp b/internal/core/src/storage/ChunkCache.cpp
index b2ddb14d63..5e1e98d600 100644
--- a/internal/core/src/storage/ChunkCache.cpp
+++ b/internal/core/src/storage/ChunkCache.cpp
@@ -106,7 +106,7 @@ ChunkCache::Read(const std::string& filepath,
                        data_type);
         } else {
             std::vector<std::shared_ptr<Chunk>> chunks{chunk};
-            column = std::make_shared<ChunkedColumn>(chunks);
+            column = std::make_shared<ChunkedColumn>(field_meta, chunks);
         }
         if (mmap_enabled && mmap_rss_not_need) {
             auto ok = madvise(reinterpret_cast<void*>(