Currently we can manipulate batch size of dataframe in 2 ways:

• DataFrame::parallelize
• DataFrame::collect

First one will split Rows into smaller batches, second one will merge small Rows into bigger batch.
Even that those two are working fine, it might not be always easy to determine which one to use, plus
their names might be confusing to users without understanding how DataFrame works.

We can solve both of those problems by adding:

DataFrame::batchSize(int $size) : self

That will split or merge processed rows based on the size.
It should also be more intuitive to the users.