[jira] [Resolved] (PHOENIX-736) How to design the table？

["Gabriel Reid (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/PHOENIX-736?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Gabriel Reid resolved PHOENIX-736.
----------------------------------

    Resolution: Fixed

Bulk resolve of closed issues imported from GitHub. This status was reached by first re-opening all closed imported issues and then resolving them in bulk.

> How to design the table？
> ------------------------
>
>                 Key: PHOENIX-736
>                 URL: https://issues.apache.org/jira/browse/PHOENIX-736
>             Project: Phoenix
>          Issue Type: Task
>            Reporter: yang ming
>
> CREATE TABLE IF NOT EXISTS YK_VIDEO_STAT (
>      DATE DATE NOT NULL,
>      VIDEOID VARCHAR NOT NULL,
>      DEVICE VARCHAR NOT NULL,
>      VV BIGINT,
>      TS BIGINT,
>      COMMENT BIGINT,
>      UP BIGINT,
>      DOWN BIGINT,
>      FAVOR BIGINT,
>      QUOTE BIGINT
>      CONSTRAINT PK PRIMARY KEY (DATE DESC, VIDEOID, DEVICE)
> ) VERSIONS=1,COMPRESSION='SNAPPY';
> this is the ddl of my table.there will have 15 million rows every day,so 3 month will
> be 1350 million. (videoid in (1~200 million interger,every day only 10 million distinct),device in ('phone','pad','pc','other'))
> Running a query similar to the following  costs at least 40s,it is too long.Can anybody give me some suggestions if query at least in the 3 month range ?
> select date,videoid,device,sum(vv),sum(ts),sum(comment),sum(up),sum(down),sum(favor),sum(quote) from yk_video_stat where videoid=10034569 and date >= to_date('20130901','yyMMdd') and date<=to_date('20130910','yyMMdd')



--
This message was sent by Atlassian JIRA
(v6.2#6252)