diff --git a/be/src/vec/core/block.cpp b/be/src/vec/core/block.cpp
index 29c6d21bb78eb78..bb37b2130c094a9 100644
--- a/be/src/vec/core/block.cpp
+++ b/be/src/vec/core/block.cpp
@@ -53,6 +53,7 @@
 #include "vec/columns/columns_number.h"
 #include "vec/common/assert_cast.h"
 #include "vec/data_types/data_type_factory.hpp"
+#include "vec/data_types/data_type_nullable.h"
 
 class SipHash;
 
@@ -476,7 +477,7 @@ std::string Block::dump_types() const {
     return out;
 }
 
-std::string Block::dump_data(size_t begin, size_t row_limit) const {
+std::string Block::dump_data(size_t begin, size_t row_limit, bool allow_null_mismatch) const {
     std::vector<std::string> headers;
     std::vector<size_t> headers_size;
     for (const auto& it : data) {
@@ -515,7 +516,14 @@ std::string Block::dump_data(size_t begin, size_t row_limit) const {
             }
             std::string s;
             if (data[i].column) {
-                s = data[i].to_string(row_num);
+                if (data[i].type->is_nullable() && !data[i].column->is_nullable()) {
+                    assert(allow_null_mismatch);
+                    s = assert_cast<const DataTypeNullable*>(data[i].type.get())
+                                ->get_nested_type()
+                                ->to_string(*data[i].column, row_num);
+                } else {
+                    s = data[i].to_string(row_num);
+                }
             }
             if (s.length() > headers_size[i]) {
                 s = s.substr(0, headers_size[i] - 3) + "...";
diff --git a/be/src/vec/core/block.h b/be/src/vec/core/block.h
index a8a4d07b49d0a0d..0055ceca23b3850 100644
--- a/be/src/vec/core/block.h
+++ b/be/src/vec/core/block.h
@@ -255,14 +255,20 @@ class Block {
 
     bool empty() const { return rows() == 0; }
 
-    /** Updates SipHash of the Block, using update method of columns.
+    /** 
+      * Updates SipHash of the Block, using update method of columns.
       * Returns hash for block, that could be used to differentiate blocks
       *  with same structure, but different data.
       */
     void update_hash(SipHash& hash) const;
 
-    /** Get block data in string. */
-    std::string dump_data(size_t begin = 0, size_t row_limit = 100) const;
+    /** 
+     *  Get block data in string. 
+     *  If code is in default_implementation_for_nulls or something likely, type and column's nullity could
+     *   temporarily be not same. set allow_null_mismatch to true to dump it correctly.
+    */
+    std::string dump_data(size_t begin = 0, size_t row_limit = 100,
+                          bool allow_null_mismatch = false) const;
 
     static std::string dump_column(ColumnPtr col, DataTypePtr type) {
         ColumnWithTypeAndName type_name {col, type, ""};
diff --git a/be/src/vec/functions/function.cpp b/be/src/vec/functions/function.cpp
index cfc6a39f397c58f..1fea4c70fc1753e 100644
--- a/be/src/vec/functions/function.cpp
+++ b/be/src/vec/functions/function.cpp
@@ -216,7 +216,8 @@ Status PreparedFunctionImpl::default_implementation_for_nulls(
         }
         RETURN_IF_ERROR(execute_without_low_cardinality_columns(context, block, new_args, result,
                                                                 block.rows(), dry_run));
-        // after run with nested, wrap them in null.
+        // After run with nested, wrap them in null. Before this, block.get_by_position(result).type
+        // is not compatible with get_by_position(result).column
         block.get_by_position(result).column = wrap_in_nullable(
                 block.get_by_position(result).column, block, args, result, input_rows_count);
 
diff --git a/fe/fe-core/src/main/java/org/apache/doris/planner/SortNode.java b/fe/fe-core/src/main/java/org/apache/doris/planner/SortNode.java
index 24b384d445357f4..8945f85ea424c14 100644
--- a/fe/fe-core/src/main/java/org/apache/doris/planner/SortNode.java
+++ b/fe/fe-core/src/main/java/org/apache/doris/planner/SortNode.java
@@ -196,6 +196,19 @@ public String getNodeExplainString(String detailPrefix, TExplainLevel detailLeve
         if (useTwoPhaseReadOpt) {
             output.append(detailPrefix + "OPT TWO PHASE\n");
         }
+
+        output.append(detailPrefix + "algorithm: ");
+        boolean isFixedLength = info.getOrderingExprs().stream().allMatch(e -> !e.getType().isStringType()
+                && !e.getType().isCollectionType());
+        if (limit > 0 && limit + offset < 1024 && (useTwoPhaseReadOpt || hasRuntimePredicate
+                || isFixedLength)) {
+            output.append("heap sort\n");
+        } else if (limit > 0 && !isFixedLength && limit + offset < 256) {
+            output.append("topn sort\n");
+        } else {
+            output.append("full sort\n");
+        }
+
         output.append(detailPrefix).append("offset: ").append(offset).append("\n");
         return output.toString();
     }