hi zeliu just want to make sure how you're updating the cuboid values in hbase? Are you merely adding the new rows to the cuboid? it might not work.
Take the cuboid (dt,code) as an example, it might already contains a row like "2016-03-01 001" : "100". Your storm outputs another row "2016-03-01 001" : "50", simply add this row to hbase might not work because currently kylin assumes all row key are distinct in a given cuboid. A safer way to approach your goal is to aggregate these two rows and overwrite the original "2016-03-01 001" : "100" to "2016-03-01 001" : "150". It seems to me that you're inventing a new way for "streaming data on Kylin platform". I suggest you discuss more of it with the community so that we could make joint efforts to promote it as a fundamental feature of Kylin. Otherwise any design changes in Kylin might break your addons, making it impossible for you to upgrade. On Fri, Mar 25, 2016 at 4:29 PM, zeLiu <[email protected]> wrote: > hi kylin team, > the version of kylin is 1.2. > I build a cube first,and every field does not use dictionary. > then,I pre calculation every cuboid in storm and update hbase table data > realtime. > > I checked the rowkey in hbase is right,but When I query data, only a few > are > correct. > > for example: > select dt,code,sum(money) as money from mytable a inner join dimtable b on > a.id=b.id group by dt,code; > 2016-03-01 001 100 > 2016-03-01 002 100 > 2016-03-02 001 200 > select dt,sum(money) as money from mytable a inner join dimtable b on > a.id=b.id group by dt; > 2016-03-01 100 > 2016-03-02 200 > > the money is 200 in 2016-03-01 ,but the result is 100, > > Why ? Is there other processing when query? > > thanks > > > > -- > View this message in context: > http://apache-kylin.74782.x6.nabble.com/update-hbase-data-realtime-and-query-it-tp3959.html > Sent from the Apache Kylin mailing list archive at Nabble.com. > -- Regards, *Bin Mahone | 马洪宾* Apache Kylin: http://kylin.io Github: https://github.com/binmahone
