[fix](round) fix round decimal128 overflow (#37733)

## Proposed changes

Issue Number: close ##37143

<!--Describe your changes.-->

be/src/vec/exec/format/format_common.h

@@ -33,7 +33,7 @@ struct DecimalScaleParams {
     int64_t scale_factor = 1;
 
     template <typename DecimalPrimitiveType>
-    static inline constexpr DecimalPrimitiveType get_scale_factor(int32_t n) {
+    static inline constexpr DecimalPrimitiveType::NativeType get_scale_factor(int32_t n) {
         if constexpr (std::is_same_v<DecimalPrimitiveType, Decimal32>) {
             return common::exp10_i32(n);
         } else if constexpr (std::is_same_v<DecimalPrimitiveType, Decimal64>) {
@@ -42,6 +42,8 @@ struct DecimalScaleParams {
             return common::exp10_i128(n);
         } else if constexpr (std::is_same_v<DecimalPrimitiveType, Decimal128V3>) {
             return common::exp10_i128(n);
+        } else if constexpr (std::is_same_v<DecimalPrimitiveType, Decimal256>) {
+            return common::exp10_i256(n);
         } else {
             static_assert(!sizeof(DecimalPrimitiveType),
                           "All types must be matched with if constexpr.");

be/src/vec/functions/round.h

@@ -32,6 +32,7 @@
 #include "vec/core/types.h"
 #include "vec/data_types/data_type.h"
 #include "vec/data_types/data_type_nullable.h"
+#include "vec/exec/format/format_common.h"
 #include "vec/functions/function.h"
 #if defined(__SSE4_1__) || defined(__aarch64__)
 #include "util/sse_util.hpp"
@@ -126,7 +127,7 @@ struct IntegerRoundingComputation {
         __builtin_unreachable();
     }
 
-    static ALWAYS_INLINE T compute(T x, T scale, size_t target_scale) {
+    static ALWAYS_INLINE T compute(T x, T scale, T target_scale) {
         switch (scale_mode) {
         case ScaleMode::Zero:
         case ScaleMode::Positive:
@@ -163,21 +164,22 @@ class DecimalRoundingImpl {
                                 Int16 out_scale) {
         Int16 scale_arg = in_scale - out_scale;
         if (scale_arg > 0) {
-            size_t scale = int_exp10(scale_arg);
+            auto scale = DecimalScaleParams::get_scale_factor<T>(scale_arg);
 
             const NativeType* __restrict p_in = reinterpret_cast<const NativeType*>(in.data());
             const NativeType* end_in = reinterpret_cast<const NativeType*>(in.data()) + in.size();
             NativeType* __restrict p_out = reinterpret_cast<NativeType*>(out.data());
 
             if (out_scale < 0) {
+                auto negative_scale = DecimalScaleParams::get_scale_factor<T>(-out_scale);
                 while (p_in < end_in) {
-                    Op::compute(p_in, scale, p_out, int_exp10(-out_scale));
+                    *p_out = Op::compute(*p_in, scale, negative_scale);
                     ++p_in;
                     ++p_out;
                 }
             } else {
                 while (p_in < end_in) {
-                    Op::compute(p_in, scale, p_out, 1);
+                    *p_out = Op::compute(*p_in, scale, 1);
                     ++p_in;
                     ++p_out;
                 }
@@ -191,11 +193,12 @@ class DecimalRoundingImpl {
                                 Int16 out_scale) {
         Int16 scale_arg = in_scale - out_scale;
         if (scale_arg > 0) {
-            size_t scale = int_exp10(scale_arg);
+            auto scale = DecimalScaleParams::get_scale_factor<T>(scale_arg);
             if (out_scale < 0) {
-                Op::compute(&in, scale, &out, int_exp10(-out_scale));
+                auto negative_scale = DecimalScaleParams::get_scale_factor<T>(-out_scale);
+                out = Op::compute(in, scale, negative_scale);
             } else {
-                Op::compute(&in, scale, &out, 1);
+                out = Op::compute(in, scale, 1);
             }
         } else {
             memcpy(&out, &in, sizeof(NativeType));

regression-test/data/nereids_p0/sql_functions/math_functions/test_round_overflow.out

@@ -0,0 +1,10 @@
+-- This file is automatically generated. You should know what you did if you want to edit this
+-- !select1 --
+186
+
+-- !select2 --
+0
+
+-- !select3 --
+20000000000000000000000
+

regression-test/suites/nereids_p0/sql_functions/math_functions/test_round_overflow.groovy

@@ -0,0 +1,25 @@
+// Licensed to the Apache Software Foundation (ASF) under one
+// or more contributor license agreements.  See the NOTICE file
+// distributed with this work for additional information
+// regarding copyright ownership.  The ASF licenses this file
+// to you under the Apache License, Version 2.0 (the
+// "License"); you may not use this file except in compliance
+// with the License.  You may obtain a copy of the License at
+//
+//   http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing,
+// software distributed under the License is distributed on an
+// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+// KIND, either express or implied.  See the License for the
+// specific language governing permissions and limitations
+// under the License.
+
+suite("test_round_overflow") {
+    sql "SET enable_nereids_planner=true"
+    sql "SET enable_fallback_to_original_planner=false"
+
+    qt_select1 "select round(coalesce(186,-33280029.8473323000000000000000000));"
+    qt_select2 "select round(coalesce(186,-33280029.8473323000000000000000000), -20);"
+    qt_select3 "select round(coalesce(18618500001234567890123.0,-33280029.847332300000000),-22);"
+}