南修子学Hbase-Hbase-Hive关联表

两个框架要关联，需要相互操作的包 我这边是cdh6.3

cp /opt/cloudera/parcels/CDH/lib/hive/lib/hive-hbase-handler.jar /opt/cloudera/parcels/CDH/lib/hbase/lib

hive 建表语句

create table hive_hbase_student(
id int,
name string,
address string) 
STORED BY 'org.apache.hadoop.hive.hbase.HBaseStorageHandler' 
WITH SERDEPROPERTIES ("hbase.columns.mapping" = ":key,info:name,info:address")
TBLPROPERTIES("hbase.table.name"="hbase_student");

在hbase插入几条语句

hbase(main):016:0> put 'hbase_student','1001','info:name','nanxiuzi'
hbase(main):011:0> put 'hbase_student','1001','info:address','jiangxi ganzhou'
hbase(main):017:0> put 'hbase_student','1002','info:name','dongdongdong'
hbase(main):014:0> put 'hbase_student','1002','info:address','jiangxi fuzhou'

在hbase查看一下

hbase(main):018:0> scan 'hbase_student'
ROW                      COLUMN+CELL                                                        
 1001                    column=info:address, timestamp=1617777531243, value=jiangxi ganzhou
 1001                    column=info:name, timestamp=1617777900745, value=nanxiuzi          
 1002                    column=info:address, timestamp=1617777564477, value=jiangxi fuzhou 
 1002                    column=info:name, timestamp=1617777911978, value=dongdongdong      
2 row(s)
Took 0.0160 seconds

在hive里面查看一下

hive> select * from hive_hbase_student;

21/04/07 14:47:20 INFO mapreduce.TableInputFormatBase: Input split length: 0 bytes.
1001    nanxiuzi        jiangxi ganzhou
1002    dongdongdong    jiangxi fuzhou

如果hbase里面表已经存在的话，需要在创建hive表的时候使用external关键字，不然会提示hbase表已经存在的错误

通过hbase进行数据插入的时候要注意类型匹配，虽然在hbase里面都是以字节存在的，但是hive里面是有类型的