westonpace commented on a change in pull request #8984:
URL: https://github.com/apache/arrow/pull/8984#discussion_r554221000
##########
File path: cpp/src/arrow/array/concatenate.cc
##########
@@ -163,6 +163,46 @@ static Status PutOffsets(const std::shared_ptr<Buffer>&
src, Offset first_offset
return Status::OK();
}
+struct DictionaryConcatenate {
+ DictionaryConcatenate(BufferVector& index_buffers,
+ std::vector<std::shared_ptr<Buffer>>& index_lookup,
+ MemoryPool* pool)
+ : out_(nullptr),
+ index_buffers_(index_buffers),
+ index_lookup_(index_lookup),
+ pool_(pool) {}
+
+ template <typename T>
+ enable_if_t<!is_integer_type<T>::value, Status> Visit(const T& t) {
+ return Status::Invalid("Dictionary indices must be integral types");
+ }
+
+ template <typename T, typename CType = typename T::c_type>
+ enable_if_integer<T, Status> Visit(const T& index_type) {
+ int64_t out_length = 0;
+ for (const auto& buffer : index_buffers_) {
+ out_length += buffer->size();
+ }
+ ARROW_ASSIGN_OR_RAISE(out_, AllocateBuffer(out_length, pool_));
+ auto out_data = out_->mutable_data();
Review comment:
Yes, the missing `reinterpret_cast` was a bug (as I think you noticed)
for index types with more than one byte. Fixed.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
