[Feature] Add option to split output into multiple files

[hugues31]

Add one or multiple options to allow user specify a strategy to split the dataset among multiple files.

It could be great for example to have :

info:
    output_name: test
    output_format: parquet
    rows: 2_000_000
    files: 5

So each file will contains approx. 2M/5 = 400k rows.

We could have parameters like:

• files : described above
• target_size : split when the file is above a certain threshold (to test HDFS optimal block size for example)
• ...