fix

be/src/vec/data_types/data_type_decimal.h

@@ -315,9 +315,16 @@ DataTypePtr decimal_result_type(const DataTypeDecimal<T>& tx, const DataTypeDeci
 
         size_t multiply_precision = tx.get_precision() + ty.get_precision();
         size_t divide_precision = tx.get_precision() + ty.get_scale();
+        // In the previous pr this was calculated as
+        // std::max(tx.get_precision() - tx.get_scale(), ty.get_precision() - ty.get_scale()) + scale + 1
+        // this is to make sure that there is space for rounding when performing add.
+        // But in fact, the decimal type has already been converted in fe, if we continue to +1 here
+        // it will result in the precision of the func return type being 1 larger than the precision of the return type generated by fe.
+        // (fe + 1 and here + 1 again, result is + 2)
+        // This will cause return_type->equals(*func_return_type) to return false when building the Function
         size_t plus_minus_precision =
                 std::max(tx.get_precision() - tx.get_scale(), ty.get_precision() - ty.get_scale()) +
-                scale + 1;
+                scale;
         if (is_multiply) {
             scale = tx.get_scale() + ty.get_scale();
             precision = std::min(multiply_precision, max_decimal_precision<Decimal256>());