lidavidm commented on a change in pull request #11257:
URL: https://github.com/apache/arrow/pull/11257#discussion_r720615425
##########
File path: cpp/src/arrow/compute/kernels/aggregate_basic.cc
##########
@@ -121,6 +122,84 @@ Result<std::unique_ptr<KernelState>>
CountInit(KernelContext*,
static_cast<const CountOptions&>(*args.options));
}
+// ----------------------------------------------------------------------
+// Distinct Count implementation
+
+template <typename Type>
+struct CountDistinctImpl : public ScalarAggregator {
+ using MemoTable = typename arrow::internal::HashTraits<Type>::MemoTableType;
+
+ explicit CountDistinctImpl(MemoryPool* memory_pool, CountOptions options)
+ : options(std::move(options)), memo_table_(new MemoTable(memory_pool,
0)) {}
+
+ Status Consume(KernelContext*, const ExecBatch& batch) override {
+ if (batch[0].is_array()) {
+ const ArrayData& arr = *batch[0].array();
+ auto visit_null = [&]() {
+ if (this->nulls > 0) return Status::OK();
+ ++this->nulls;
+ return Status::OK();
+ };
+ auto visit_value = [&](typename Type::c_type arg) {
+ int y;
+ RETURN_NOT_OK(memo_table_->GetOrInsert(arg, &y));
+ return Status::OK();
+ };
+ RETURN_NOT_OK(VisitArrayDataInline<Type>(arr, visit_value, visit_null));
+ this->non_nulls += this->memo_table_->size();
+ } else {
+ const Scalar& input = *batch[0].scalar();
+ this->nulls += !input.is_valid * batch.length;
+ this->non_nulls += input.is_valid * batch.length;
+ }
Review comment:
I think that comment is misleading/incorrect based on how the query
engine/scanner have developed. You could have an ExecBatch formed by projecting
just the partition columns from a batch, in which case you'd have scalar rows
with a length > 1.
--
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]
