Reply
Highlighted
Contributor
Posts: 47
Registered: ‎07-27-2015

Hive on Hbase with strange behavior when group by etc.

HI,

  We are working with kerberos CDH 5.7.3 & CM 5.8. 

  I create a Hive Table on HBase with the below command:

 

create external table arch_mr_jobs
(job_id STRING,
dt STRING,
a STRING,
b STRING,
.......
)STORED BY 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'
("hbase.columns.mapping"=":key,d:dt,d:a,d:b,......) TBLPROPERTIES("hbase.table.name"="arch:mr_jobs);

 

There is just one row that has the column d:a in hbase table arch:mr_jobs. There are no d:a column in other rows.  So, the strange behavior comming:

select a,count(1) from arch_mr_jobs;  got:
FAILED   |    1
--You can see: There is just one row ,I hope the result should be:
FAILED   |     1
NULL      |      50
--why the null value be ignore? This is a wrong result.

select a from arch_mr_jobs;  got:
FAILED   |   
--I hope the result should be:
FAILED
NULL
NULL
.
.

 

 So, I believe I missed some config. I google but got nothing. 

Could you give me any point?

Thank you at advance

Paul