Backport #92076 to 25.12: Add additional checks for missing streams in Wide parts

robot-clickhouse · robot-clickhouse · commit 17cffcad0b0c · 2026-01-19T11:17:54.000Z
diff --git a/src/DataTypes/Serializations/SerializationNullable.cpp b/src/DataTypes/Serializations/SerializationNullable.cpp
@@ -22,6 +22,7 @@ namespace ErrorCodes
 {
     extern const int CANNOT_READ_ALL_DATA;
     extern const int INCORRECT_DATA;
+    extern const int LOGICAL_ERROR;
 }
 
 void SerializationNullable::enumerateStreams(
@@ -159,7 +160,7 @@ void SerializationNullable::deserializeBinaryBulkWithMultipleStreams(
     auto nested_column = col.getNestedColumnPtr();
     if (null_map->size() != nested_column->size())
         throw Exception(
-            ErrorCodes::INCORRECT_DATA,
+            settings.native_format ? ErrorCodes::INCORRECT_DATA : ErrorCodes::LOGICAL_ERROR,
             "Sizes of nested column and null map of Nullable column are not equal after deserialization (null map size = {}, nested "
             "column size = {})",
             null_map->size(),
diff --git a/src/DataTypes/Serializations/SerializationObject.cpp b/src/DataTypes/Serializations/SerializationObject.cpp
@@ -1048,6 +1048,20 @@ void SerializationObject::deserializeBinaryBulkWithMultipleStreams(
     settings.path.pop_back();
     settings.path.pop_back();
 
+    /// Verify that all typed paths, dynamic paths and shared data has consistent sizes
+    size_t expected_size = shared_data->size();
+    for (const auto & [path, path_column] : typed_paths)
+    {
+        if (path_column->size() != expected_size)
+            throw Exception(settings.native_format ? ErrorCodes::INCORRECT_DATA : ErrorCodes::LOGICAL_ERROR, "Unexpected size of typed path {}: {}. Expected size {}", path, path_column->size(), expected_size);
+    }
+
+    for (const auto & [path, path_column] : dynamic_paths)
+    {
+        if (path_column->size() != expected_size)
+            throw Exception(settings.native_format ? ErrorCodes::INCORRECT_DATA : ErrorCodes::LOGICAL_ERROR, "Unexpected size of dynamic path {}: {}. Expected size {}", path, path_column->size(), expected_size);
+    }
+
     column_object.repairDuplicatesInDynamicPathsAndSharedData(shared_data_previous_size);
 }
 
diff --git a/src/DataTypes/Serializations/SerializationTuple.cpp b/src/DataTypes/Serializations/SerializationTuple.cpp
@@ -19,6 +19,7 @@ namespace ErrorCodes
     extern const int SIZES_OF_COLUMNS_IN_TUPLE_DOESNT_MATCH;
     extern const int NOT_FOUND_COLUMN_IN_BLOCK;
     extern const int INCORRECT_DATA;
+    extern const int LOGICAL_ERROR;
 }
 
 
@@ -820,6 +821,14 @@ void SerializationTuple::deserializeBinaryBulkWithMultipleStreams(
             column_tuple.getColumnPtr(i), rows_offset, limit, settings, tuple_state->states[i], cache);
     }
 
+    /// Verify that all Tuple elements have the same size.
+    size_t expected_size = column_tuple.getColumn(0).size();
+    for (size_t i = 1; i < elems.size(); ++i)
+    {
+        if (column_tuple.getColumn(i).size() != expected_size)
+            throw Exception(settings.native_format ? ErrorCodes::INCORRECT_DATA : ErrorCodes::LOGICAL_ERROR, "Unexpected size of tuple element {}: {}. Expected size: {}", i, column_tuple.getColumn(i).size(), expected_size);
+    }
+
     typeid_cast<ColumnTuple &>(*mutable_column).addSize(column_tuple.getColumn(0).size());
 }
 
diff --git a/src/DataTypes/Serializations/SerializationVariant.cpp b/src/DataTypes/Serializations/SerializationVariant.cpp
@@ -607,6 +607,10 @@ void SerializationVariant::deserializeBinaryBulkWithMultipleStreams(
             col.getVariantPtrByLocalDiscriminator(i), variant_rows_offsets[i], variant_limits[i],
             settings, variant_state->variant_states[i], cache);
         settings.path.pop_back();
+
+        /// Verify that we deserialized data of this variant.
+        if (variant_limits[i] && col.getVariantPtrByLocalDiscriminator(i)->empty())
+            throw Exception(settings.native_format ? ErrorCodes::INCORRECT_DATA : ErrorCodes::LOGICAL_ERROR, "Variant {} is empty, but expected to be read {} values", variant_names[i], variant_limits[i]);
     }
     settings.path.pop_back();
 
diff --git a/src/Storages/MergeTree/MergeTreeReaderWide.cpp b/src/Storages/MergeTree/MergeTreeReaderWide.cpp
@@ -24,6 +24,11 @@ namespace
     constexpr auto DATA_FILE_EXTENSION = ".bin";
 }
 
+namespace ErrorCodes
+{
+    extern const int LOGICAL_ERROR;
+}
+
 MergeTreeReaderWide::MergeTreeReaderWide(
     MergeTreeDataPartInfoForReaderPtr data_part_info_,
     NamesAndTypesList columns_,
@@ -327,7 +332,22 @@ ReadBuffer * MergeTreeReaderWide::getStream(
 
     auto stream_name = IMergeTreeDataPart::getStreamNameForColumn(name_and_type, substream_path, ".bin", checksums, storage_settings);
     if (!stream_name)
+    {
+        /// We allow missing streams only for columns/subcolumns that are not present in this part.
+        auto column = data_part_info_for_read->getColumnsDescription().tryGetColumn(GetColumnsOptions::AllPhysical, name_and_type.getNameInStorage());
+        if (column && (!name_and_type.isSubcolumn() || column->type->hasSubcolumn(name_and_type.getSubcolumnName())))
+        {
+            throw Exception(
+                ErrorCodes::LOGICAL_ERROR,
+                "Stream {} for column {} with type {} is not found",
+                ISerialization::getFileNameForStream(
+                    name_and_type.type->getName(), substream_path, ISerialization::StreamFileNameSettings(*storage_settings)),
+                    name_and_type.name,
+                    column->type->getName());
+        }
+
         return nullptr;
+    }
 
     auto it = streams.find(*stream_name);
     if (it == streams.end())

Original file line number	Diff line number	Diff line change
`@@ -19,6 +19,7 @@ namespace ErrorCodes`
`19`	`19`	`extern const int SIZES_OF_COLUMNS_IN_TUPLE_DOESNT_MATCH;`
`20`	`20`	`extern const int NOT_FOUND_COLUMN_IN_BLOCK;`
`21`	`21`	`extern const int INCORRECT_DATA;`
	`22`	`+ extern const int LOGICAL_ERROR;`
`22`	`23`	`}`
`23`	`24`
`24`	`25`
`@@ -820,6 +821,14 @@ void SerializationTuple::deserializeBinaryBulkWithMultipleStreams(`
`820`	`821`	`column_tuple.getColumnPtr(i), rows_offset, limit, settings, tuple_state->states[i], cache);`
`821`	`822`	`}`
`822`	`823`
	`824`	`+ /// Verify that all Tuple elements have the same size.`
	`825`	`+ size_t expected_size = column_tuple.getColumn(0).size();`
	`826`	`+ for (size_t i = 1; i < elems.size(); ++i)`
	`827`	`+ {`
	`828`	`+ if (column_tuple.getColumn(i).size() != expected_size)`
	`829`	`+ throw Exception(settings.native_format ? ErrorCodes::INCORRECT_DATA : ErrorCodes::LOGICAL_ERROR, "Unexpected size of tuple element {}: {}. Expected size: {}", i, column_tuple.getColumn(i).size(), expected_size);`
	`830`	`+ }`
	`831`	`+`
`823`	`832`	`typeid_cast<ColumnTuple &>(*mutable_column).addSize(column_tuple.getColumn(0).size());`
`824`	`833`	`}`
`825`	`834`