Method BuildTree

cpp/src/parquet/schema.cc:830–857

}

void SchemaDescriptor::BuildTree(const NodePtr& node, int16_t max_def_level,
                                 int16_t max_rep_level, const NodePtr& base) {
  if (node->is_optional()) {
    ++max_def_level;
  } else if (node->is_repeated()) {
    // Repeated fields add a definition level. This is used to distinguish
    // between an empty list and a list with an item in it.
    ++max_rep_level;
    ++max_def_level;
  }

  // Now, walk the schema and create a ColumnDescriptor for each leaf node
  if (node->is_group()) {
    const GroupNode* group = static_cast<const GroupNode*>(node.get());
    for (int i = 0; i < group->field_count(); ++i) {
      BuildTree(group->field(i), max_def_level, max_rep_level, base);
    }
  } else {
    node_to_leaf_index_[static_cast<const PrimitiveNode*>(node.get())] =
        static_cast<int>(leaves_.size());

    // Primitive node, append to leaves
    leaves_.push_back(ColumnDescriptor(node, max_def_level, max_rep_level, this));
    leaf_to_base_.emplace(static_cast<int>(leaves_.size()) - 1, base);
    leaf_to_idx_.emplace(node->path()->ToDotString(),
                         static_cast<int>(leaves_.size()) - 1);
  }
}

int SchemaDescriptor::GetColumnIndex(const PrimitiveNode& node) const {
  auto it = node_to_leaf_index_.find(&node);