Hi Fred,
I ran your benchmark and I get similar results. That is, I can't
duplicate hm:j's findings that going to the raw connection is slower.
So +1 from me, and as James H has done, I've put an alternate patch
there - use select_value rather than select_all.
Trevor
On 22-Dec-07, at 8:46 AM, Frederick Cheung wrote:
>
>
> On 21 Dec 2007, at 18:48, Frederick Cheung wrote:
>
>>
>> On 21 Dec 2007, at 18:18, Josh Susser wrote:
>>
>>>
>>>
>>> I'm the author of the new exists? code, and I also looked at
>>> doing it
>>> with a lower-level connection call instead of a find with :select
>>> option. I don't have the data around anymore, but as I recall, the
>>> benchmarks showed that using the select_all approach didn't perform
>>> as
>>> well, so I went with the find with :select approach. I hadn't
>>> thought
>>> about the after_find callback interaction, and I guess no one else
>>> has
>>> run into that problem before now either. Some stuff has changed in
>>> AR
>>> that might have affected performance of the select_all approach
>>> since
>>> I tried it last, so it might be worth revisiting now.
>>>
>>
>> Well I ended up with that code by taking apart find :first and
>> removing the instantiation of rows at the end, so I don't think it
>> should be a problem.
>> What I posted should be doing exactly what find :first does, but
>> without the instantiation
>
> I ran some numbers today:
>
> ids = (1..10000).map { rand 2000000}
>
> Benchmark.bmbm do |x|
> x.report("exists") { ids.each {|id| Question.exists? id} }
> end
>
> Current trunk (find:first)
>
> Rehearsal ------------------------------------------
> exists 4.700000 0.620000 5.320000 ( 6.437581)
> --------------------------------- total: 5.320000sec
>
> user system total real
> exists 4.690000 0.620000 5.310000 ( 6.574966)
>
> select_all
>
> Rehearsal ------------------------------------------
> exists 4.650000 0.740000 5.390000 ( 8.984423)
> --------------------------------- total: 5.390000sec
>
> user system total real
> exists 4.010000 0.590000 4.600000 ( 5.606526)
>
>
> Not mind blowing and it's obviously not going to make a massive
> difference to anyone, but for me at least sets aside any performance
> worries.
>
> Fred
>
>
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups "Ruby
on Rails: Core" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/rubyonrails-core?hl=en
-~----------~----~----~----~------~----~------~--~---
