No, I couldn't really get any traction on this so I let it slide. The 
consensus on the Github issue was to leave the `count` inconsistency as-is. 
The last word from jonleighton was:

 @avit <https://github.com/avit> coming back to this, I am feeling less 
> confident that anything needs to be changed in AR.
>
> We already have a way to create subquery via #from and you can then 
> perform a count on the subquery if you wish. This to me seems sufficient to 
> handle what is a rare case in most applications.
>
> So if I understand correctly, the remaining objection is that #countmight 
> return a Fixnum or a Hash depending on the situation. I probably 
> wouldn't have designed it like that in the first place, but given it's 
> there I'm not sure there is a strong argument for changing it...
>
 I still disagree and would've loved to work on a solution before Rails 4, 
but that's where it's at.

Andrew Vit



On Thursday, July 11, 2013 12:55:43 PM UTC-7, Kyle Shipley wrote:
>
> Did this ever get patched? I'm running into this issue currently and would 
> love to know if it got merged anywhere.
>
> On Friday, October 12, 2012 3:02:41 AM UTC-4, dburry wrote:
>>
>> There have been times when I would have liked to be able to somehow 
>> consistently get a fixnum count, instead of sometimes a fixnum and 
>> sometimes a hash that I just have to loop through and add up the values!
>>
>> Dave
>>
>>
>> On Thursday, October 11, 2012 7:01:47 PM UTC-7, Andrew Vit wrote:
>>>
>>> I posted my proposed changes to the Calculations API, and resulting SQL 
>>> queries here:
>>>
>>> https://gist.github.com/3681183#file_calculations_b.rb
>>>
>>> I'm preferring the "calculations_b.rb" version over the one with a 
>>> separate grouped_count method: it would be a less intrusive change, and 
>>> when we need to pass a :group option anyway, it doesn't make sense to 
>>> repeat `grouped_count(group:...)` when `count(group:...)` is obvious.
>>>
>>> Andrew Vit
>>>
>>> On Thursday, October 11, 2012 6:55:48 PM UTC-7, Andrew Vit wrote:
>>>>
>>>> I'd like to get some traction or feedback on issue #7121 so I know if 
>>>> it's worth me doing it. What are the objections to:
>>>>
>>>> 1. Fixing brittle calculation methods via subqueries?
>>>> 2. Requiring count(:group=>:author_id) option explicitly when we want a 
>>>> Hash instead of Fixnum?
>>>>
>>>> Or am I the only one who thinks this is broken? :-(
>>>>
>>>> Andrew Vit
>>>>
>>>

-- 
You received this message because you are subscribed to the Google Groups "Ruby 
on Rails: Core" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/rubyonrails-core.
For more options, visit https://groups.google.com/groups/opt_out.


Reply via email to