[Developers] mqlread queries and caching

[Jason Douglas]

Having no mwLastWriteTime cookie does not avoid caching... pretty much  
the opposite.  The only way you to avoid caching is to present a  
"current" mwLastWriteTime cookie obtained in the response to an upload  
or write or by calling the touch service.

-jason


On Sep 8, 2008, at 1:35 PM, Augusto Callejas wrote:

> wouldn't it be less work (one less roundtrip HTTP request) to clear  
> out the
> value of "mwLastWriteTime" instead of calling "/api/service/touch"?   
> i want
> to do a series of queries that avoid caching.
>
> thanks
> augusto
>
>
> On 9/8/08 11:21 AM, "Ben Drees" <ben at metaweb.com> wrote:
>
>> That's correct - You must track cookies for /api/service/touch to  
>> have
>> an effect.
>>
>> An individual user agent that tracks cookies will automatically  
>> read its
>> own writes.
>>
>> An individual user agent that does not track cookies may not always
>> automatically read its own writes.
>>
>> /api/service/touch is useful when one user agent needs to read the
>> results of another user agent's recent writes.
>>
>> The cache-suppressing HTTP request headers work, but may result in
>> poorer performance than with the /api/service/touch approach.
>>
>> David Flanagan wrote:
>>> Doesn't /api/service/touch work by simply returning a  
>>> nwLastWriteTime
>>> cookie?  So if you don't track the cookie in the first place, then  
>>> the
>>> touch service won't do you any good, if I understand correctly.
>>>
>>> David
>>>
>>> John Giannandrea wrote:
>>>
>>>> Augusto,
>>>> If its the same program doing the writes and the reads you should  
>>>> just
>>>> track the mwLastWriteTime cookie rather than calling /api/service/
>>>> touch every time.
>>>> -jg
>>>> On Sep 5, 2008, at 4:41 PM, Phil Gochenour wrote:
>>>>
>>>>> Hi Augusto,
>>>>>
>>>>> have a look at this Help topic and let us know if it solves the
>>>>> problem for you:
>>>>>
>>>>> Avoid caching problems when using the API
>>>>> http://www.freebase.com/view/guid/9202a8c04000641f80000000076224e8
>>>>>
>>>>> Phil
>>>>>
>>>> _______________________________________________
>>>> Developers mailing list
>>>> Developers at freebase.com
>>>> http://lists.freebase.com/mailman/listinfo/developers
>>>>
>>>>
>>>
>>> _______________________________________________
>>> Developers mailing list
>>> Developers at freebase.com
>>> http://lists.freebase.com/mailman/listinfo/developers
>>>
>>>
>>
>> _______________________________________________
>> Developers mailing list
>> Developers at freebase.com
>> http://lists.freebase.com/mailman/listinfo/developers
>
>
> _______________________________________________
> Developers mailing list
> Developers at freebase.com
> http://lists.freebase.com/mailman/listinfo/developers