[ 
https://issues.apache.org/jira/browse/ARROW-11030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17256105#comment-17256105
 ] 

Andy Grove commented on ARROW-11030:
------------------------------------

I have a theory on what might be happening here but I am struggling to really 
understand this.

It looks like we create a buffer and for each input array, we extend this 
buffer. Each time we extend it, the buffer is larger so the cost of extending 
it again gets higher each time?

Is there a way we can compute upfront how much to extend it by and do one 
extend operation?

> [Rust] [DataFusion] Poor join performance with smaller batches
> --------------------------------------------------------------
>
>                 Key: ARROW-11030
>                 URL: https://issues.apache.org/jira/browse/ARROW-11030
>             Project: Apache Arrow
>          Issue Type: Bug
>          Components: Rust - DataFusion
>            Reporter: Andy Grove
>            Assignee: Andy Grove
>            Priority: Major
>             Fix For: 3.0.0
>
>
> Performance of joins slows down dramatically with smaller batches.
> The issue is related to slow performance of MutableDataArray::new() when 
> passed a high number of batches. This happens when passing in all of the 
> batches from the build side of the join and this happens once per build-side 
> join key for each probe-side batch.
> It seems to get exponentially slower as the number of arrays increases even 
> though the number of rows is the same.
> I modified hash_join.rs to have this debug code:
> {code:java}
> let start = Instant::now();
> let row_count: usize = arrays.iter().map(|arr| arr.len()).sum();
> let num_arrays = arrays.len();
> let mut mutable = MutableArrayData::new(arrays, true, capacity);
> if num_arrays > 0 {
>     debug!("MutableArrayData::new() with {} arrays containing {} rows took {} 
> ms", num_arrays, row_count, start.elapsed().as_millis());
> } {code}
> Batch size 131072:
> {code:java}
> MutableArrayData::new() with 4584 arrays containing 3115341 rows took 1 ms
> MutableArrayData::new() with 4584 arrays containing 3115341 rows took 1 ms
> MutableArrayData::new() with 4584 arrays containing 3115341 rows took 1 ms 
> {code}
> Batch size 16384:
> {code:java}
> MutableArrayData::new() with 36624 arrays containing 3115341 rows took 19 ms
> MutableArrayData::new() with 36624 arrays containing 3115341 rows took 16 ms
> MutableArrayData::new() with 36624 arrays containing 3115341 rows took 17 ms 
> {code}
> Batch size 4096:
> {code:java}
> MutableArrayData::new() with 146496 arrays containing 3115341 rows took 88 ms
> MutableArrayData::new() with 146496 arrays containing 3115341 rows took 89 ms
> MutableArrayData::new() with 146496 arrays containing 3115341 rows took 88 ms 
> {code}
>  
>  
>  
>  
>  



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to