Is your feature request related to a problem or challenge? Please describe what you are trying to do.

If some process creates a large number of parquet files, it would be nice to be able to run parquet-concat on these (location: parquet/src/bin/parquet-concat.rs). The current script will however open up too many file descriptors on most systems in a situation of many files.

Describe the solution you'd like

Make the implementation streaming based: files are opened, read and written from in order. The current implementation starts by opening all file descriptors.

Describe alternatives you've considered

Additional context