Function ExecNode_Scan

r/src/compute-exec.cpp:278–309

[[dataset::export]]

// [[dataset::export]]
std::shared_ptr<acero::ExecNode> ExecNode_Scan(
    const std::shared_ptr<acero::ExecPlan>& plan,
    const std::shared_ptr<ds::Dataset>& dataset,
    const std::shared_ptr<compute::Expression>& filter, cpp11::list projection) {
  arrow::dataset::internal::Initialize();

  // TODO: pass in FragmentScanOptions
  auto options = std::make_shared<ds::ScanOptions>();

  options->use_threads = arrow::r::GetBoolOption("arrow.use_threads", true);
  options->dataset_schema = dataset->schema();

  // This filter is only used for predicate pushdown;
  // you still need to pass it to a FilterNode after to handle any other components
  options->filter = *filter;

  // ScanNode needs to know which fields to materialize.
  // It will pull them from this projection to prune the scan,
  // but you still need to Project after
  std::vector<compute::Expression> exprs;
  for (SEXP expr : projection) {
    auto expr_ptr = cpp11::as_cpp<std::shared_ptr<compute::Expression>>(expr);
    exprs.push_back(*expr_ptr);
  }
  cpp11::strings field_names(projection.attr(R_NamesSymbol));
  options->projection = call(
      "make_struct", std::move(exprs),
      compute::MakeStructOptions{cpp11::as_cpp<std::vector<std::string>>(field_names)});

  return MakeExecNodeOrStop("scan", plan.get(), {},
                            ds::ScanNodeOptions{dataset, options});
}

// [[dataset::export]]
void ExecPlan_Write(