Apache Kylin1.5.2.1之订单案例详细构建流程

转：http://blog.itpub.net/30089851/viewspace-2122586/

一.Hive订单数据仓库构建

1. 创建事实表并插入数据

DROP TABLE IF EXISTS default.fact_order ;
create table default.fact_order (
　　time_key string,
　　product_key string,
　　salesperson_key string,
　　custom_key string,
　　quantity_ordered bigint,
　　order_dollars bigint,
　　cost_dollars bigint
)
ROW FORMAT DELIMITED FIELDS TERMINATED BY ','
STORED AS TEXTFILE;

load data local inpath '/root/kylinsample/fact_order.txt' overwrite into table default.fact_order;

##load data local inpath '/root/kylinsample/fact_order.txt'  into table default.fact_order;

fact_order.txt

2016-05-01,pd001,sp001,ct001,100,2000,1000
2016-05-01,pd001,sp002,ct002,100,2000,1000
2016-05-01,pd001,sp003,ct002,100,2000,1000
2016-05-01,pd002,sp002,ct002,100,2000,1000
2016-05-01,pd003,sp003,ct001,100,2000,1000
2016-05-01,pd001,sp003,ct001,100,2000,1000
2016-05-01,pd001,sp002,ct001,100,2000,1000
2016-05-01,pd001,sp003,ct002,100,2000,1000
2016-05-01,pd002,sp001,ct001,100,2000,1000
2016-05-01,pd003,sp001,ct001,100,2000,1000
2016-05-01,pd004,sp001,ct001,50,1000,600
2016-05-02,pd001,sp001,ct001,50,1000,600
2016-05-02,pd001,sp002,ct002,100,2000,1000
2016-05-02,pd001,sp003,ct002,100,2000,1000
2016-05-02,pd002,sp001,ct001,50,1000,600
2016-05-02,pd003,sp001,ct001,50,1000,600
2016-05-02,pd004,sp001,ct001,50,1000,600
2016-05-03,pd001,sp001,ct001,50,1000,600
2016-05-03,pd001,sp002,ct002,100,2000,1000
2016-05-03,pd001,sp003,ct002,100,2000,1000
2016-05-04,pd002,sp001,ct001,700,14000,10000
2016-05-04,pd003,sp001,ct001,700,14000,10000
2016-05-04,pd004,sp001,ct001,100,2000,1000
2016-05-05,pd001,sp001,ct001,100,2000,1000
2016-05-05,pd001,sp002,ct002,700,14000,10000
2016-05-05,pd001,sp003,ct002,700,14000,10000
2016-05-05,pd002,sp001,ct001,100,2000,1000
2016-05-05,pd003,sp001,ct001,100,2000,1000
2016-05-05,pd004,sp001,ct001,100,2000,1000
2016-05-06,pd001,sp001,ct001,100,2000,1000
2016-05-06,pd001,sp002,ct002,100,2000,1000
2016-05-06,pd001,sp003,ct002,100,2000,1000
2016-05-07,pd002,sp001,ct001,100,2000,1000
2016-05-07,pd003,sp001,ct001,100,2000,1000
2016-05-07,pd004,sp001,ct001,50,1000,600
2016-05-07,pd002,sp001,ct001,100,2000,1000
2016-05-07,pd003,sp001,ct001,100,2000,1000
2016-05-07,pd004,sp001,ct001,50,1000,600
2016-05-08,pd001,sp001,ct001,50,1000,600
2016-05-08,pd001,sp002,ct002,100,2000,1000
2016-05-08,pd001,sp003,ct002,100,2000,1000
2016-05-08,pd001,sp001,ct001,50,1000,600
2016-05-08,pd001,sp002,ct002,100,2000,1000
2016-05-08,pd001,sp003,ct002,100,2000,1000
2016-05-08,pd001,sp001,ct001,50,1000,600
2016-05-08,pd001,sp002,ct002,100,2000,1000
2016-05-08,pd001,sp003,ct002,100,2000,1000
2016-05-09,pd002,sp001,ct001,50,1000,600
2016-05-09,pd003,sp001,ct001,50,1000,600
2016-05-09,pd004,sp001,ct001,50,1000,600
2016-05-09,pd001,sp001,ct001,50,1000,600
2016-05-09,pd002,sp001,ct001,50,1000,600
2016-05-09,pd003,sp001,ct001,50,1000,600
2016-05-09,pd004,sp001,ct001,50,1000,600
2016-05-09,pd001,sp001,ct001,50,1000,600
2016-05-09,pd001,sp002,ct002,100,2000,1000
2016-05-09,pd004,sp003,ct002,100,2000,1000
2016-05-09,pd002,sp001,ct001,700,14000,10000
2016-05-09,pd003,sp003,ct001,700,14000,10000
2016-05-09,pd004,sp003,ct001,100,2000,1000
2016-05-10,pd001,sp001,ct001,100,2000,1000
2016-05-10,pd001,sp002,ct002,700,14000,10000
2016-05-10,pd001,sp003,ct002,700,14000,10000
2016-05-10,pd002,sp001,ct001,100,2000,1000
2016-05-11,pd003,sp003,ct001,100,2000,1000
2016-05-11,pd004,sp001,ct001,100,2000,1000
2016-05-12,pd001,sp001,ct001,100,2000,1000
2016-05-12,pd004,sp002,ct002,100,2000,1000
2016-05-12,pd001,sp003,ct002,100,2000,1000
2016-05-12,pd001,sp001,ct001,100,2000,1000
2016-05-12,pd004,sp002,ct002,100,2000,1000
2016-05-12,pd001,sp003,ct002,100,2000,1000
2016-05-13,pd002,sp001,ct001,100,2000,1000
2016-05-13,pd003,sp001,ct001,100,2000,1000
2016-05-13,pd004,sp001,ct001,50,1000,600
2016-05-14,pd001,sp001,ct001,50,1000,600
2016-05-14,pd001,sp002,ct002,100,2000,1000
2016-05-14,pd001,sp003,ct002,100,2000,1000
2016-05-15,pd002,sp001,ct001,50,1000,600
2016-05-15,pd003,sp001,ct001,50,1000,600
2016-05-15,pd004,sp001,ct001,50,1000,600
2016-05-15,pd002,sp001,ct001,50,1000,600
2016-05-15,pd003,sp001,ct001,50,1000,600
2016-05-15,pd004,sp001,ct001,50,1000,600
2016-05-15,pd002,sp001,ct001,50,1000,600
2016-05-15,pd003,sp001,ct001,50,1000,600
2016-05-15,pd004,sp001,ct001,50,1000,600
2016-05-16,pd001,sp001,ct001,50,1000,600
2016-05-16,pd001,sp002,ct002,100,2000,1000
2016-05-16,pd001,sp003,ct002,100,2000,1000
2016-05-16,pd001,sp001,ct001,50,1000,600
2016-05-16,pd001,sp002,ct002,100,2000,1000
2016-05-16,pd001,sp003,ct002,100,2000,1000
2016-05-17,pd002,sp001,ct001,700,14000,10000
2016-05-17,pd003,sp001,ct001,700,14000,10000
2016-05-17,pd004,sp001,ct001,100,2000,1000
2016-05-17,pd002,sp001,ct001,700,14000,10000
2016-05-17,pd003,sp001,ct001,700,14000,10000
2016-05-17,pd004,sp001,ct001,100,2000,1000
2016-05-18,pd001,sp001,ct001,100,2000,1000
2016-05-18,pd003,sp002,ct001,700,14000,10000
2016-05-18,pd001,sp003,ct002,700,14000,10000
2016-05-19,pd002,sp001,ct001,100,2000,1000
2016-05-19,pd003,sp001,ct002,100,2000,1000
2016-05-20,pd001,sp001,ct001,100,2000,1000
2016-05-20,pd002,sp002,ct002,100,2000,1000
2016-05-20,pd003,sp003,ct001,100,2000,1000
2016-05-20,pd004,sp001,ct001,100,2000,1000
2016-05-20,pd001,sp002,ct002,100,2000,1000
2016-05-20,pd002,sp001,ct002,100,2000,1000

 
2. 创建天维度表dim_day

DROP TABLE IF EXISTS default.dim_day ;

create table default.dim_day (
　　day_key string,
　　full_day string,
　　month_name string,
　　quarter string,
　　year string
)
ROW FORMAT DELIMITED FIELDS TERMINATED BY ','
STORED AS TEXTFILE; 
load data local inpath '/root/kylinsample/dim_day.txt' overwrite into table default.dim_day;

dim_day.txt
  
2016-05-01,2016-05-01,201605,2016q2,2016
2016-05-02,2016-05-02,201605,2016q2,2016
2016-05-03,2016-05-03,201605,2016q2,2016
2016-05-04,2016-05-04,201605,2016q2,2016
2016-05-05,2016-05-05,201605,2016q2,2016
2016-05-06,2016-05-06,201605,2016q2,2016
2016-05-07,2016-05-07,201605,2016q2,2016
2016-05-08,2016-05-08,201605,2016q2,2016
2016-05-09,2016-05-09,201605,2016q2,2016
2016-05-10,2016-05-10,201605,2016q2,2016
2016-05-11,2016-05-11,201605,2016q2,2016
2016-05-12,2016-05-12,201605,2016q2,2016
2016-05-13,2016-05-13,201605,2016q2,2016
2016-05-14,2016-05-14,201605,2016q2,2016
2016-05-15,2016-05-15,201605,2016q2,2016
2016-05-16,2016-05-16,201605,2016q2,2016
2016-05-17,2016-05-17,201605,2016q2,2016
2016-05-18,2016-05-18,201605,2016q2,2016
2016-05-19,2016-05-19,201605,2016q2,2016
2016-05-20,2016-05-20,201605,2016q2,2016

  
3. 创建售卖员的维度表salesperson_dim
 
DROP TABLE IF EXISTS default.dim_salesperson ;
 
create table default.dim_salesperson (
　　salesperson_key string,
　　salesperson string,
　　salesperson_id string,
　　region string,
　　region_code string
)
ROW FORMAT DELIMITED FIELDS TERMINATED BY ','
STORED AS TEXTFILE;
  
load data local inpath '/root/kylinsample/dim_salesperson.txt' overwrite into table default.dim_salesperson;
  
dim_salesperson.txt
  
sp001,hongbin,sp001,beijing,10086
sp002,hongming,sp002,beijing,10086
sp003,hongmei,sp003,beijing,10086

4. 创建客户维度 custom_dim

 
DROP TABLE IF EXISTS default.dim_custom ;
  
create table default.dim_custom (
　　custom_key string,
　　custom_name string,
　　custorm_id string,
　　headquarter_states string,
　　billing_address string,
　　billing_city string,
　　billing_state string,
　　industry_name string
)
ROW FORMAT DELIMITED FIELDS TERMINATED BY ','
STORED AS TEXTFILE;
 
load data local inpath '/root/kylinsample/dim_custom.txt' overwrite into table default.dim_custom;

dim_custom.txt
  
ct001,custom_john,ct001,beijing,zgx-beijing,beijing,beijing,internet                    
ct002,custom_herry,ct002,henan,shlinjie,shangdang,henan,internet     
 
 
5. 创建产品维度表并插入数据
 
DROP TABLE IF EXISTS default.dim_product ;                                              
                                                                                          
create table default.dim_product (                                                      
　　product_key string,                                                                 
　　product_name string,                                                                
　　product_id string,                                                                  
　　product_desc string,                                                                
　　sku string,                                                                         
　　brand string,                                                                       
　　brand_code string,                                                                  
　　brand_manager string,                                                               
　　category string,                                                                    
category_code string                                                                
)                                                                                       
ROW FORMAT DELIMITED FIELDS TERMINATED BY ','                                           
STORED AS TEXTFILE;                                                                     
                       
load data local inpath '/root/kylinsample/dim_product.txt' overwrite into table default.dim_product;      
dim_product.txt
  
pd001,Box-Large,pd001,Box-Large-des,large1.0,brand001,brandcode001,brandmanager001,Packing,cate001
pd002,Box-Medium,pd001,Box-Medium-des,medium1.0,brand001,brandcode001,brandmanager001,Packing,cate001
pd003,Box-small,pd001,Box-small-des,small1.0,brand001,brandcode001,brandmanager001,Packing,cate001
pd004,Evelope,pd001,Evelope_des,large3.0,brand001,brandcode001,brandmanager001,Pens,cate002

 
这样一个星型的结构表在hive中创建完毕, 实际上一个离线的数据仓库已经完成, 它包含一个主题, 即商品订单.

三.Kylin的Project创建与数据同步
　　1.单击"Manage Project" 
　　2.单击"New Project"
　　3.输入"Project Name", WareHouse_01
　　4.Submit

　　1.选择WareHouse_01,选择"Data Source" tab页
　　2.单击"Load Hive Table"
　　3.输入需要同步的表
  　 　　"DEFAULT.FACT_ORDER,DEFAULT.DIM_DAY,DEFAULT.DIM_PRODUCT,DEFAULT.DIM_SALESPERSON,DEFAULT.DIM_CUSTOM"
　　4.Sync

四.Kylin的Model创建
　　1.选择"Models" tab页,单击"New Model"
　　2."Model Name"输入,WareHouse_01_Model
　　3.选择"Fact Table"为 DEFAULT.FACT_ORDER;再 添加Lookup Table;
　　4.选取每张表的哪些列字段作为Dimensions
 　　　　ID Table Name           Columns
 　　　　1.DEFAULT.FACT_ORDER  TIME_KEY PRODUCT_KEY SALESPERSON_KEY CUSTOM_KEY
 　　　　2.DEFAULT.DIM_DAY          FULL_DAY
 　　　　3.DEFAULT.DIM_PRODUCT  PRODUCT_NAME
 　　　　4.DEFAULT.DIM_SALESPERSON  SALESPERSON
 　　　　5.DEFAULT.DIM_CUSTOM  CUSTOM_NAME

　　5.选取DEFAULT.FACT_ORDER表的哪些列字段作为measures
      　　QUANTITY_ORDERED ORDER_DOLLARS COST_DOLLARS

　　6.a.选取 "Partition Date Column"为DEFAULT.FACT_ORDER.TIME_KEY,格式 yyyy-MM-dd
　　　　b.对于"Filter"条件,由于没有要过滤的条件,故不填写

　　7.Save

五.Kylin的Cube创建

　　1.选择"Models" tab页,单击"New Cube“

　　2.Cube Info:
      　　"Model Name"选择,WareHouse_01_Model
      　　"Cube Name"输入,cube01

　　3.Dismensions:
      　　单击"Auto Generator",依据情况选择维度的列,全选

　　4.Measures:
　　　　a.单击"+Measure",添加要聚合计算的度量,比如 sum(QUANTITY_ORDERED)
　　　　b.Expression: SUM/MIN/MAX/COUNT/COUNT_DISTINCT/TOP_N/RAW
　　5.Refresh Setting:
　　　　a.Auto Merge Thresholds,自动合并阈值,7~28 days
　　　　b.Retention Threshold,保留天数,60
　　　　c.Partition Start Date,非常重要,是后面build cube的开始日期

　　6.Advanced Setting:
　　　　--Aggregation Groups:
　　　　a.Includes: TIME_KEY ,PRODUCT_KEY ,SALESPERSON_KEY , CUSTOM_KEY
　　　　b.Mandatory Dimensions: TIME_KEY
　　　　c.Hierarchy Dimensions: PRODUCT_KEY ,SALESPERSON_KEY ,CUSTOM_KEY
　　　　d.Joint Dimensions: 无
　　　　
　　　　--Rowkeys:
　　　　TIME_KEY ,PRODUCT_KEY ,SALESPERSON_KEY ,CUSTOM_KEY 4个字段为dict字典编码
 
　　7.Configuration Overwrites: 无

　　8.Overview:
　　　　保存cube

五.Cube Build

　　1.选择 cube01,单击”Action”,选择Build

　　2.填写End Date,Submit

　　3.单击”Monitor”,观察Job

　　4.等待Process100% (Any Errors)

　　5.返回cube01,查看 cube size 和 Source Records等字段更新

 
六.Hive* kyin 查询对比

　　点击(此处)折叠或打开

1.2016-05-01到2016-05-15期间的每天的订单数量,订单金额,订单成本

Hive: 65.816 s
select fact.time_key, sum(fact.quantity_ordered), sum(fact.order_dollars), sum(fact.cost_dollars) from fact_order as fact
where fact.time_key >= "2016-05-01" and fact.time_key <= "2016-05-15"
group by fact.time_key order by fact.time_key;

Kylin: 0.32s-->0.27s
select fact.time_key, sum(fact.quantity_ordered), sum(fact.order_dollars), sum(fact.cost_dollars) from fact_order as fact
where fact.time_key between '2016-05-01' and '2016-05-15'
group by fact.time_key order by fact.time_key

　　点击(此处)折叠或打开

2.2016-05-01到2016-05-15期间的每天的产品的订单量

Hive: 100.336s
select dday.full_day,dsp.product_name, sum(fact.quantity_ordered) from fact_order as fact
inner join dim_day as dday on fact.time_key = dday.day_key
inner join dim_product as dsp on fact.product_key = dsp.product_key
where dday.full_day >= "2016-05-01" and dday.full_day <= "2016-05-15"
group by dday.full_day,dsp.product_name
order by dday.full_day,dsp.product_name;

Kylin:0.93s-->0.39s
select dday.full_day,dsp.product_name, sum(fact.quantity_ordered) from fact_order as fact
inner join dim_day as dday on fact.time_key = dday.day_key
inner join dim_product as dsp on fact.product_key = dsp.product_key
where dday.full_day >= '2016-05-01' and dday.full_day <= '2016-05-15'
group by dday.full_day,dsp.product_name
order by dday.full_day,dsp.product_name