dhruv9vats commented on a change in pull request #12484:
URL: https://github.com/apache/arrow/pull/12484#discussion_r823847455
##########
File path: cpp/src/arrow/compute/kernels/hash_aggregate.cc
##########
@@ -2758,6 +2770,336 @@ struct GroupedOneFactory {
InputType argument_type;
};
+// ----------------------------------------------------------------------
+// List implementation
+
+template <typename Type, typename Enable = void>
+struct GroupedListImpl final : public GroupedAggregator {
+ using CType = typename TypeTraits<Type>::CType;
+ using GetSet = GroupedValueTraits<Type>;
+
+ Status Init(ExecContext* ctx, const std::vector<ValueDescr>&,
+ const FunctionOptions* options) override {
+ ctx_ = ctx;
+ has_nulls_ = false;
+ // out_type_ initialized by GroupedListInit
+ values_ = TypedBufferBuilder<CType>(ctx_->memory_pool());
+ groups_ = TypedBufferBuilder<uint32_t>(ctx_->memory_pool());
+ values_bitmap_ = TypedBufferBuilder<bool>(ctx_->memory_pool());
+ return Status::OK();
+ }
+
+ Status Resize(int64_t new_num_groups) override {
+ num_groups_ = new_num_groups;
+ return Status::OK();
+ }
+
+ Status Consume(const ExecBatch& batch) override {
+ // DCHECK_EQ(batch[0].array()->offset, 0);
+ const auto* groups = batch[1].array()->GetValues<uint32_t>(1);
+ const auto* values = batch[0].array()->GetValues<CType>(1);
+ int64_t num_values = batch[0].array()->length;
+
+ RETURN_NOT_OK(groups_.Append(groups, num_values));
+ RETURN_NOT_OK(GetSet::AppendBuffers(values_, values, num_values));
+
+ if (batch[0].null_count() > 0) {
+ if (!has_nulls_) {
+ has_nulls_ = true;
+ RETURN_NOT_OK(values_bitmap_.Append(num_args_, true));
+ }
+ const auto* values_bitmap = batch[0].array()->GetValues<uint8_t>(0);
+ RETURN_NOT_OK(values_bitmap_.Reserve(num_values));
+ values_bitmap_.UnsafeAppend(values_bitmap, 0, num_values);
Review comment:
Does the current structure look somewhat correct? Also, please suggest a
more rigorous way of testing.
--
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: github-unsubscr...@arrow.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
