Struct SmallFilePlanner 

pub struct SmallFilePlanner { /* private fields */ }

Plans scan parallelism for a set of files.

When individual files are smaller than target_bytes, multiple files are grouped into a single task so each task processes roughly target_bytes of data. Files larger than target_bytes each get their own task (splitting within a file is not yet supported).

Implementations

impl SmallFilePlanner

pub fn new(target_bytes: u64) -> Self

Create a planner with the given target bytes per task.

pub fn plan(&self, files: &[FileStats]) -> SplitPlanAdvice

Produce a scan plan for the given file list.

Files are grouped greedily: accumulate until the next file would push the group over target_bytes, then start a new group. This ensures each group is at most target_bytes + max_single_file_bytes.