Fewer bytes scanned in single node plans

[jayshrivastava]

See this comment / test https://github.com/datafusion-contrib/datafusion-distributed/pull/307/changes#r2718771809

In certain queries, the file scan is less efficient when distributing. We fail to prune data based on statistics.

One theory as to why this happens is because we mess up the statistics when creating new file groups in scale_up_leaf_node implementation for FileScanConfig.

This needs investigation.

[gabotechs]

One theory as to why this happens is because we mess up the statistics when creating new file groups in scale_up_leaf_node implementation for FileScanConfig.

🤔 This is a good theory, it would be nice to investigate that