1.hbase数据关联hive外部表

create external table hive_hbase_test1(id string,address string,age int,gender string,name string)

stored by 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'

with serdeproperties("hbase.columns.mapping"=":key,info:address,info:age,info:gender,info:name")

tblproperties("hbase.table.name"="hbase_table");

2.hive beeline导出到本地文件

beeline -u jdbc:hive2://192.168.21.110:2166 -d org.apache.hive.jdbc.HiveDriver --showHeader=false --outputformat=txt -e "select * from hive_hbase_test1" > f2.txt

showHeader=false 不要头部，纯净模式

outputformat=txt 文件格式

f2.txt文件用 | 隔开

3.本地文件上传hdfs

hadoop fs -put /home/user/f2.txt /tmp/

4.hdfs导入hive

先创建表（注意指定表的数据格式）STORED AS TEXTFILE

create table test_2000(id string,collect_time string,create_time string,status string,value string)

row format delimited fields terminated by '|'

STORED AS INPUTFORMAT 'org.apache.hadoop.mapred.TextInputFormat'

OUTPUTFORMAT 'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat'

load命令导入hive

beeline -u jdbc:hive2://192.168.1.110:2066 -e "load data inpath '/home/f2.txt' into table test_2000;"

hive表数据格式参考：https://blog.csdn.net/TC_HaoShuai/article/details/84303140

5.hive创建中间表关联habse

CREATE TABLE hive_hbase_test_2000(id string,collect_time string,create_time string,status string,value string)

STORED BY 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'

WITH SERDEPROPERTIES ("hbase.columns.mapping" = ":key,info:address,info:age,info:gender,info:name")

TBLPROPERTIES ("hbase.table.name" = "test_2000",

"hbase.mapred.output.outputtable" = "test_2000");

6.hive原数据往hive中间表导入

insert into table hive_hbase_test_2000 select * from test_2000;

查看hbase表也有了数据 scan 'test_2000' ，但是这种情况删除外部表hbase表也会没有

7.避免hbase表与hive强关联

可以先创建hbase表

create 'test_2000','info'

然后hive外部关联hbase表，

CREATE external TABLE hive_hbase_test_2000(id string,collect_time string,create_time string,status string,value string)