[ 
https://issues.apache.org/jira/browse/CALCITE-4522?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17296425#comment-17296425
 ] 

hqx edited comment on CALCITE-4522 at 3/10/21, 3:50 AM:
--------------------------------------------------------

Thanks, I add some test and make the issue description more detailed. I suggest 
the follow formula(for sort case, no sort just return 0) because we can use 
heap sort if offset + fetch < input_count, else we can use heap sort, quick 
sort and so on.

sort_cpu_cost  = log(min(offset + fetch, input_count)) * input_count* row_byte

 


was (Author: 871):
Thanks, I add some test and make the issue description more detailed. I suggest 
the follow formula(for sort case, no sort is just 0) because we can use heap 
sort if offset + fetch < input_count, else we can use heap sort, quick sort and 
so on.

sort_cpu_cost  = log(min(offset + fetch, input_count)) * input_count* row_byte

 

> Sort operator returns the same cpu cost no matter the RelCollation is empty 
> or not
> ----------------------------------------------------------------------------------
>
>                 Key: CALCITE-4522
>                 URL: https://issues.apache.org/jira/browse/CALCITE-4522
>             Project: Calcite
>          Issue Type: Improvement
>          Components: core
>            Reporter: hqx
>            Priority: Minor
>              Labels: pull-request-available
>          Time Spent: 40m
>  Remaining Estimate: 0h
>
> The old method to compute the cost of sort has some problem.
>  # When the RelCollation is empty, there is no need to sort, but it still 
> compute the cpu cost of sort.
>  # use n * log\(n) * row_byte to estimate the cpu cost may be inaccurate, 
> where n means the output row count of the sort operator, and row_byte means 
> the average bytes of one row .
> Instead, I give follow suggestion.
>  # the cpu cost is zero if the RelCollation is empty.
>  # let heap_size be min\(offset + output_count, input_count), and use 
> input_count * log\(heap_size)* row_byte to compute the cpu cost.
> When fetch is zero, I found the output_count is 1 not 0. This conveniently 
> ensure the log\(heap_size) no less than zero



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

Reply via email to