a small optimization here is to use `std::vector<std::shared_ptr<Table>> tables(row_group.size(), nullptr)` above and use `&tables[i]` here. Afaik certain tools like Spark generate files with many small RowGroups.
[ Full content available at: https://github.com/apache/parquet-cpp/pull/492 ] This message was relayed via gitbox.apache.org for [email protected]
