phoenix-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ravi Kiran <maghamraviki...@gmail.com>
Subject Re: explain plan -parallel X-way
Date Wed, 12 Mar 2014 14:32:09 GMT
Hi Yosi,

   You can have a look at DefaultParallelIteratorRegionSplitter.java code
to understand  how the api creates splits based on various heuristics like
phoenix.query.targetConcurrency and phoenix.query.maxConcurrency .
   The default value of phoenix.query.targetConcurrency is 32.  Try
changing the value for targetConcurrency to 31 and I believe you would see
31 parallel scans.

  Please note , I haven't tested this .

Regards
Ravi


On Wed, Mar 12, 2014 at 7:14 PM, Yosi Botzer <yosi.botzer@gmail.com> wrote:

> Hi,
>
> There is something I cannot understand regrading the explain plan of
> phoenix. I have a table that has 31 regions all together, however the
> explain plan of the query below suggests that there are 62 parallel scans,
>
> how could it be?
>
> Yosi
>
>
> 0: jdbc:phoenix:ec2-54-203-158-70.us-west-2.c> explain select campaign_id,
> count(*) from fact5 group by campaign_id;
> +------------+
> |    PLAN    |
> +------------+
> | CLIENT PARALLEL *62-WAY* FULL SCAN OVER FACT5 |
> |     SERVER FILTER BY FIRST KEY ONLY |
> |     SERVER AGGREGATE INTO DISTINCT ROWS BY [CAMPAIGN_ID] |
> | CLIENT MERGE SO
>

Mime
View raw message