[GitHub] [arrow] andygrove commented on a change in pull request #7958: ARROW-9725: [Rust] [DataFusion] SortExec and LimitExec re-use MergeExec

[GitBox <gi...@apache.org>]

andygrove commented on a change in pull request #7958:
URL: https://github.com/apache/arrow/pull/7958#discussion_r470182745



##########
File path: rust/datafusion/src/execution/physical_plan/limit.rs
##########
@@ -81,39 +82,40 @@ struct LimitPartition {
 
 impl Partition for LimitPartition {
     fn execute(&self) -> Result<Arc<Mutex<dyn RecordBatchReader + Send + Sync>>> {
-        // collect up to "limit" rows on each partition
-        let threads: Vec<JoinHandle<Result<Vec<RecordBatch>>>> = self
+        // apply limit in parallel across all input partitions
+        let local_limit = self
             .partitions
             .iter()
             .map(|p| {
-                let p = p.clone();
-                let limit = self.limit;
-                thread::spawn(move || {
-                    let it = p.execute()?;
-                    collect_with_limit(it, limit)
-                })
+                Arc::new(LocalLimitExec::new(

Review comment:
       I agree. I would like us to move in that direction. It would be worth filing a JIRA for that as an improvement. Would you like to do that?




----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org