Paimon blob 4.1

be/src/core/data_type/data_type_file.h

@@ -31,6 +31,11 @@ class DataTypeFile final : public IDataType {
 public:
     static constexpr PrimitiveType PType = TYPE_FILE;
     static constexpr size_t FIELD_COUNT = 6;
+    static constexpr size_t URI_FIELD_INDEX = 0;
+    static constexpr size_t OFFSET_FIELD_INDEX = 1;
+    static constexpr size_t SIZE_FIELD_INDEX = 2;
+    static constexpr size_t CONTENT_TYPE_FIELD_INDEX = 3;
+    static constexpr size_t CHECKSUM_FIELD_INDEX = 4;
     static constexpr size_t INLINE_FIELD_INDEX = 5;
 
     DataTypeFile();

be/src/exec/scan/access_path_parser.cpp

@@ -439,6 +439,10 @@ Status build_nested_children_from_access_node(format::ColumnDefinition* column,
     }
 
     switch (nested_type->get_primitive_type()) {
+    case TYPE_FILE:
+        // FILE has a fixed six-child schema. Keep the whole value in the scan projection;
+        // element_at selects its public field after the reader materializes the FILE column.
+        return Status::OK();
     case TYPE_STRUCT:
         return build_struct_children_from_access_node(
                 column, assert_cast<const DataTypeStruct&>(*nested_type), node, path, schema_column,

be/src/exec/scan/file_scanner.cpp

@@ -44,6 +44,7 @@
 #include "core/column/column_nullable.h"
 #include "core/column/column_vector.h"
 #include "core/data_type/data_type.h"
+#include "core/data_type/data_type_file.h"
 #include "core/data_type/data_type_nullable.h"
 #include "core/data_type/data_type_string.h"
 #include "core/string_ref.h"
@@ -1055,6 +1056,16 @@ Status FileScanner::_get_next_reader() {
         const TFileRangeDesc& range = _current_range;
         _current_range_path = range.path;
 
+        if (range.__isset.table_format_params && range.table_format_params.__isset.paimon_params &&
+            (range.table_format_params.paimon_params.__isset.blob_files ||
+             std::ranges::any_of(_file_slot_descs, [](const SlotDescriptor* slot) {
+                 return contains_file_type(slot->get_data_type_ptr());
+             }))) {
+            return Status::NotSupported(
+                    "Paimon BLOB columns require enable_file_scanner_v2=true for both native "
+                    "and JNI readers");
+        }
+
         if (!_partition_slot_descs.empty()) {
             // we need get partition columns first for runtime filter partition pruning
             RETURN_IF_ERROR(_generate_partition_columns());

be/src/exprs/function/cast/cast_to_variant.h

@@ -180,8 +180,8 @@ struct CastToVariant {
 };
 
 // create corresponding variant value to wrap from_type
-WrapperType create_cast_to_variant_wrapper(const DataTypePtr& from_type,
-                                           const DataTypeVariant& to_type) {
+inline WrapperType create_cast_to_variant_wrapper(const DataTypePtr& from_type,
+                                                  const DataTypeVariant& to_type) {
     if (from_type->get_primitive_type() == TYPE_VARIANT) {
         // variant_max_subcolumns_count is not equal
         return create_unsupport_wrapper(from_type->get_name(), to_type.get_name());
@@ -190,8 +190,8 @@ WrapperType create_cast_to_variant_wrapper(const DataTypePtr& from_type,
 }
 
 // create corresponding type convert from variant
-WrapperType create_cast_from_variant_wrapper(const DataTypeVariant& from_type,
-                                             const DataTypePtr& to_type) {
+inline WrapperType create_cast_from_variant_wrapper(const DataTypeVariant& from_type,
+                                                    const DataTypePtr& to_type) {
     if (to_type->get_primitive_type() == TYPE_VARIANT) {
         // variant_max_subcolumns_count is not equal
         return create_unsupport_wrapper(from_type.get_name(), to_type->get_name());

be/src/format/jni/jni_data_bridge.cpp

@@ -24,13 +24,15 @@
 
 #include "core/block/block.h"
 #include "core/column/column_array.h"
+#include "core/column/column_file.h"
 #include "core/column/column_map.h"
 #include "core/column/column_nullable.h"
 #include "core/column/column_string.h"
 #include "core/column/column_struct.h"
 #include "core/column/column_varbinary.h"
 #include "core/column/variant_v2/column_variant_v2.h"
 #include "core/data_type/data_type_array.h"
+#include "core/data_type/data_type_file.h"
 #include "core/data_type/data_type_map.h"
 #include "core/data_type/data_type_nullable.h"
 #include "core/data_type/data_type_struct.h"
@@ -146,6 +148,15 @@ Status JniDataBridge::fill_column(TableMetaAddress& address, ColumnPtr& doris_co
     case PrimitiveType::TYPE_STRUCT:
         status = _fill_struct_column(address, data_column, data_type, num_rows);
         break;
+    case PrimitiveType::TYPE_FILE: {
+        auto& file = assert_cast<ColumnFile&>(*data_column);
+        const auto& file_type = assert_cast<const DataTypeFile&>(*remove_nullable(data_type));
+        for (size_t i = 0; i < DataTypeFile::FIELD_COUNT && status.ok(); ++i) {
+            status = fill_column(address, file.get_column_ptr(i), file_type.get_element(i),
+                                 num_rows);
+        }
+        break;
+    }
     case PrimitiveType::TYPE_VARBINARY:
         status = _fill_varbinary_column(address, data_column, num_rows);
         break;
@@ -297,6 +308,8 @@ std::string JniDataBridge::get_jni_type(const DataTypePtr& data_type) {
     DataTypePtr type = remove_nullable(data_type);
     std::ostringstream buffer;
     switch (type->get_primitive_type()) {
+    case TYPE_FILE:
+        return "file";
     case TYPE_BOOLEAN:
         return "boolean";
     case TYPE_TINYINT:
@@ -403,6 +416,8 @@ std::string JniDataBridge::get_jni_type_with_different_string(const DataTypePtr&
     DataTypePtr type = remove_nullable(data_type);
     std::ostringstream buffer;
     switch (data_type->get_primitive_type()) {
+    case TYPE_FILE:
+        return "file";
     case TYPE_BOOLEAN:
         return "boolean";
     case TYPE_TINYINT:
@@ -535,6 +550,8 @@ std::string JniDataBridge::encode_string_list(const std::vector<std::string>& va
 
 std::string JniDataBridge::get_jni_type_with_encoded_struct_fields(const DataTypePtr& data_type) {
     switch (data_type->get_primitive_type()) {
+    case TYPE_FILE:
+        return "file";
     case TYPE_STRUCT: {
         const auto* type_struct =
                 assert_cast<const DataTypeStruct*>(remove_nullable(data_type).get());
@@ -634,6 +651,15 @@ Status JniDataBridge::_fill_column_meta(const ColumnPtr& doris_column, const Dat
         }
         break;
     }
+    case PrimitiveType::TYPE_FILE: {
+        const auto& file = assert_cast<const ColumnFile&>(*data_column);
+        const auto& file_type = assert_cast<const DataTypeFile&>(*remove_nullable(data_type));
+        for (size_t i = 0; i < DataTypeFile::FIELD_COUNT; ++i) {
+            RETURN_IF_ERROR(
+                    _fill_column_meta(file.get_column_ptr(i), file_type.get_element(i), meta_data));
+        }
+        break;
+    }
     case PrimitiveType::TYPE_MAP: {
         const auto& map = assert_cast<const ColumnMap&>(*data_column);
         const auto& key_type =

be/src/format/jni/jni_data_bridge.h

@@ -130,7 +130,8 @@ class JniDataBridge {
 
     /**
      * Convert a Doris DataType to its JNI type string representation.
-     * e.g., TYPE_INT -> "int", TYPE_DECIMAL128I -> "decimal128(p,s)"
+     * e.g., TYPE_INT -> "int", TYPE_DECIMAL128I -> "decimal128(p,s)", TYPE_FILE -> "file".
+     * FILE retains its logical type; both sides use the canonical six-child physical layout.
      */
     static std::string get_jni_type(const DataTypePtr& data_type);