Currently, the CSV reader always start from 0 and go until the end, which makes it impossible for DataFusion to split a single file in parts, e.g. via multiple readers.
The goal of this issue is to generalize the reader to support starting from an arbitrary location (via iterating), and also have an optional upper limit.
Reporter: Jorge Leitão / @jorgecarleitao
Assignee: Jorge Leitão / @jorgecarleitao
PRs and other links:
Note: This issue was originally created as ARROW-10332. Please see the migration documentation for further details.