asterixdb-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Xikui Wang <xik...@uci.edu>
Subject Re: OptimizedHybridHashJoinOperatorDescriptor vs. HybridHashJoinOperatorDescriptor
Date Fri, 09 Nov 2018 07:49:01 GMT
Hi Yingyi,

Thanks for your reply. I think you are right. The two seem to serve the
same purpose except that the unoptimized one has less documentation and no
memory management. I will propose a patch later to remove that.

Best,
Xikui

On Thu, Nov 8, 2018 at 9:54 PM Yingyi Bu <buyingyi@gmail.com> wrote:

> I'm not sure if it's still correct, but based on my understanding,
> OptimizedHybridHashJoinOperatorDescriptor does the role reversal
> optimization which was done by Pouria, while
> HybridHashJoinOperatorDescriptor was the old implementation before Pouria's
> work and probably could be deleted.
>
> Best,
> Yingyi
>
> On Thu, Nov 8, 2018 at 6:52 PM Xikui Wang <xikuiw@uci.edu> wrote:
>
> > Hi Devs,
> >
> > Does anyone know what's the difference between
> > the OptimizedHybridHashJoinOperatorDescriptor and the
> > HybridHashJoinOperatorDescriptor?
> >
> > I was going over part of the join codes and found both of these classes
> > there. It seems the only difference between the two is when there is a
> > hashFunctionFamily is null, we will use the
> > HybridHashJoinOperatorDescriptor. However, it seems this would not
> > happen... In fact, I changed to code to always use the optimized one and
> no
> > test cases fail (could be test cases bias though). Thanks ahead!
> >
> > Best,
> > Xikui
> >
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message