Oh, I'm using 2.2 (will release soon).
Then you can run "bin/mestore.sh backup", and then find the model json file
in "model_desc" folder, in it you will see. After modify the
partition class, you can restore from that folder back to HBase, and then
restart Kylin.
2017-10-23 18:00 GMT+08:00 op <520075...@qq.com>:
> on kylin web,there is no "Edit Json" ......... my version 2.0
>
>
> ------------------ 原始邮件 ------------------
> *发件人:* "ShaoFeng Shi";<shaofeng...@apache.org>;
> *发送时间:* 2017年10月23日(星期一) 下午5:53
> *收件人:* "dev"<dev@kylin.apache.org>;
> *主题:* Re: data format when building increment cube
>
> You can backup the metastore to local, then you will see the model json
> file.
>
> On Kylin web, move mouse to the model, click "Action" -> "Edit JSON", will
> see the JSON content.
>
> 2017-10-23 17:23 GMT+08:00 op <520075...@qq.com>:
>
> > thank you shaofeng，but i cannot find the json file
> > or you mean change it on kylin web？
> >
> >
> >
> >
> > ------------------ 原始邮件 ------------------
> > 发件人: "ShaoFeng Shi";<shaofeng...@apache.org>;
> > 发送时间: 2017年10月23日(星期一) 下午5:09
> > 收件人: "dev"<dev@kylin.apache.org>;
> >
> > 主题: Re: data format when building increment cube
> >
> >
> >
> > If you create a new model from Kylin web, the model has a property
> > "partition_condition_builder" which refers to the class that
> > implements the PartitionConditionBuilder.
> > You can manually change it to your customized implementation:
> >
> >
> > {
> > "uuid" : "ad40c5a2-a32f-4f27-84cd-85ea43af2307",
> > "last_modified" : 1508748727953,
> > "name" : "test_parition",
> > "owner" : "admin",
> > "is_draft" : false,
> > "description" : "",
> > "fact_table" : "DEFAULT.KYLIN_SALES",
> > "lookups" : [ ],
> > "dimensions" : [ {
> > "table" : "KYLIN_SALES",
> > "columns" : [ "TRANS_ID", "PART_DT", "LSTG_FORMAT_NAME",
> > "LEAF_CATEG_ID", "LSTG_SITE_ID", "SLR_SEGMENT_CD", "SELLER_ID",
> "BUYER_ID",
> > "OPS_USER_ID", "OPS_REGION" ]
> > } ],
> > "metrics" : [ "KYLIN_SALES.PRICE", "KYLIN_SALES.ITEM_COUNT" ],
> > "filter_condition" : "",
> > "partition_desc" : {
> > "partition_date_column" : "KYLIN_SALES.PART_DT",
> > "partition_time_column" : null,
> > "partition_date_start" : 0,
> > "partition_date_format" : "yyyy-MM-dd",
> > "partition_time_format" : "HH:mm:ss",
> > "partition_type" : "APPEND",
> > "partition_condition_builder" :
> > "org.apache.kylin.metadata.model.PartitionDesc$
> > DefaultPartitionConditionBuilder"
> > },
> > "capacity" : "MEDIUM"
> > }
> >
> > 2017-10-23 16:19 GMT+08:00 op <520075...@qq.com>:
> >
> > > hi shaofeng
> > > then how to use a new PartitionConditionBuilder？
> > > thanks！
> > >
> > >
> > > ------------------ 原始邮件 ------------------
> > > 发件人: "ShaoFeng Shi";<shaofeng...@apache.org>;
> > > 发送时间: 2017年10月17日(星期二) 晚上6:16
> > > 收件人: "dev"<dev@kylin.apache.org>;
> > >
> > > 主题: Re: data format when building increment cube
> > >
> > >
> > >
> > > It is not supported. You need to develop a new
> PartitionConditionBuilder.
> > > Please refer
> > > to org.apache.kylin.metadata.model.PartitionDesc.
> > > YearMonthDayPartitionConditionBuilder
> > >
> > > 2017-10-17 18:06 GMT+08:00 op <520075...@qq.com>:
> > >
> > > >
> > > > hello everyone.
> > > > how can i define my own data format?
> > > > our hive tables are partitioned by month_id(format :yyyymm) and
> > > > subpartitioned by day_id(format ：dd)
> > > >
> > > > thanks!
> > > >
> > >
> > >
> > >
> > > --
> > > Best regards,
> > >
> > > Shaofeng Shi 史少锋
> > >
> >
> >
> >
> > --
> > Best regards,
> >
> > Shaofeng Shi 史少锋
> >
>
>
>
> --
> Best regards,
>
> Shaofeng Shi 史少锋
>
>
--
Best regards,
Shaofeng Shi 史少锋
