You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@hive.apache.org by "Raghunath, Ranjith" <Ra...@usaa.com> on 2012/05/23 02:43:07 UTC

Map side aggregations

I have the parameter hive.map.aggr set to true. However, when I look at the counters associated with the map tasks I notice the following "Combine input records 0". I am interpreting this as a failure to perform the map side aggregation. Is that accurate? Is this option not working in hive 0.7.1?

Thanks,
Ranjith

Re: Map side aggregations

Posted by Ranjith <ra...@gmail.com>.

Thanks philip.

Thanks,
Ranjith

On May 23, 2012, at 4:15 AM, Philip Tromans <ph...@gmail.com> wrote:

> Hi Ranjith,
> 
> I haven't checked the code (so this might not be true), but I think that the map side aggregation stuff uses it's own hash map within the map phase to do the aggregation, instead of using a combiner, so you wouldn't expect to see any combine input records. Have a look for parameters like hive.groupby.mapaggr.checkinterval, and the associated documentation will explain how it all works.
> 
> Cheers,
> 
> Phil.
> 
> On 23 May 2012 02:44, Ranjith <ra...@gmail.com> wrote:
> Thanks Matt. I am not performing a join so does that matter? What does this local task do?
> 
> Thanks,
> Ranjith
> 
> On May 22, 2012, at 8:17 PM, "Tucker, Matt" <Ma...@disney.com> wrote:
> 
>> Try setting hive.auto.convert.join to true.  The CLI will have a local task before it starts a map-reduce job on the cluster.
>> 
>> Matt
>> 
>> 
>> 
>> On May 22, 2012, at 8:43 PM, "Raghunath, Ranjith" <Ra...@usaa.com> wrote:
>> 
>>> I have the parameter hive.map.aggr set to true. However, when I look at the counters associated with the map tasks I notice the following “Combine input records 0”. I am interpreting this as a failure to perform the map side aggregation. Is that accurate? Is this option not working in hive 0.7.1?
>>>  
>>> Thanks,
>>> Ranjith
>>>  
>>>  
>>>  
>

Re: Map side aggregations

Posted by Philip Tromans <ph...@gmail.com>.

Hi Ranjith,

I haven't checked the code (so this might not be true), but I think that
the map side aggregation stuff uses it's own hash map within the map phase
to do the aggregation, instead of using a combiner, so you wouldn't expect
to see any combine input records. Have a look for parameters
like hive.groupby.mapaggr.checkinterval, and the associated documentation
will explain how it all works.

Cheers,

Phil.

On 23 May 2012 02:44, Ranjith <ra...@gmail.com> wrote:

> Thanks Matt. I am not performing a join so does that matter? What does
> this local task do?
>
> Thanks,
> Ranjith
>
> On May 22, 2012, at 8:17 PM, "Tucker, Matt" <Ma...@disney.com>
> wrote:
>
> Try setting hive.auto.convert.join to true.  The CLI will have a local
> task before it starts a map-reduce job on the cluster.
>
> Matt
>
>
>
> On May 22, 2012, at 8:43 PM, "Raghunath, Ranjith" <
> Ranjith.Raghunath1@usaa.com> wrote:
>
>  I have the parameter hive.map.aggr set to true. However, when I look at
> the counters associated with the map tasks I notice the following “Combine
> input records 0”. I am interpreting this as a failure to perform the map
> side aggregation. Is that accurate? Is this option not working in hive
> 0.7.1?
>
> Thanks,
> Ranjith
>
>
>
>
>

Re: Map side aggregations

Posted by Ranjith <ra...@gmail.com>.

Thanks Matt. I am not performing a join so does that matter? What does this local task do?

Thanks,
Ranjith

On May 22, 2012, at 8:17 PM, "Tucker, Matt" <Ma...@disney.com> wrote:

> Try setting hive.auto.convert.join to true.  The CLI will have a local task before it starts a map-reduce job on the cluster.
> 
> Matt
> 
> 
> 
> On May 22, 2012, at 8:43 PM, "Raghunath, Ranjith" <Ra...@usaa.com> wrote:
> 
>> I have the parameter hive.map.aggr set to true. However, when I look at the counters associated with the map tasks I notice the following “Combine input records 0”. I am interpreting this as a failure to perform the map side aggregation. Is that accurate? Is this option not working in hive 0.7.1?
>>  
>> Thanks,
>> Ranjith
>>  
>>  
>>

Re: Map side aggregations

Posted by "Tucker, Matt" <Ma...@disney.com>.

Try setting hive.auto.convert.join to true.  The CLI will have a local task before it starts a map-reduce job on the cluster.

Matt



On May 22, 2012, at 8:43 PM, "Raghunath, Ranjith" <Ra...@usaa.com>> wrote:

I have the parameter hive.map.aggr set to true. However, when I look at the counters associated with the map tasks I notice the following “Combine input records 0”. I am interpreting this as a failure to perform the map side aggregation. Is that accurate? Is this option not working in hive 0.7.1?

Thanks,
Ranjith