yiguolei commented on code in PR #52676:
URL: https://github.com/apache/doris/pull/52676#discussion_r2184117414
##########
be/src/vec/data_types/serde/data_type_number_serde.cpp:
##########
@@ -424,16 +424,38 @@ Status DataTypeNumberSerDe<T>::write_column_to_orc(const
std::string& timezone,
if constexpr (T == TYPE_LARGEINT) { // largeint
auto* cur_batch = dynamic_cast<orc::StringVectorBatch*>(orc_col_batch);
-
- INIT_MEMORY_FOR_ORC_WRITER()
-
+ // First pass: calculate total memory needed and collect serialized
values
+ size_t total_size = 0;
for (size_t row_id = start; row_id < end; row_id++) {
if (cur_batch->notNull[row_id] == 1) {
std::string value_str = fmt::format("{}", col_data[row_id]);
size_t len = value_str.size();
-
- REALLOC_MEMORY_FOR_ORC_WRITER()
-
+ total_size += len;
+ }
+ }
+ // Allocate continues memory based on calculated size
+ char* ptr = (char*)malloc(total_size);
+ if (!ptr) {
+ return Status::InternalError(
+ "malloc memory {} error when write variant column data to
orc file.",
+ total_size);
+ }
+ StringRef bufferRef;
+ bufferRef.data = ptr;
+ bufferRef.size = total_size;
+ buffer_list.emplace_back(bufferRef);
+ // Second pass: fill the data and update the batch
+ size_t offset = 0;
+ for (size_t row_id = start; row_id < end; row_id++) {
+ if (cur_batch->notNull[row_id] == 1) {
+ std::string value_str = fmt::format("{}", col_data[row_id]);
+ size_t len = value_str.size();
+ if (offset + len > total_size) {
+ return Status::InternalError(
+ "Buffer overflow when writing column data to ORC
file. offset {} with "
+ "len {} exceed total_size {} . ",
+ offset, len, total_size);
+ }
strcpy(const_cast<char*>(bufferRef.data) + offset,
value_str.c_str());
Review Comment:
把这里直接换成memcpy 吧
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
