update query in hive/hbase - hadoop

Question

I have already created a table in hbase using hive:
hive> CREATE TABLE hbase_table_emp(id int, name string, role string)
STORED BY 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'
WITH SERDEPROPERTIES ("hbase.columns.mapping" = ":key,cf1:name,cf1:role")
TBLPROPERTIES ("hbase.table.name" = "emp");
and created another table to load data on it :
hive> create table testemp(id int, name string, role string) row format delimited fields terminated by '\t';
hive> load data local inpath '/home/user/sample.txt' into table testemp;
and finally insert data into the hbase table:
hive> insert overwrite table hbase_table_emp select * from testemp;
hive> select * from hbase_table_emp;
OK
123 Ram TeamLead
456 Silva Member
789 Krishna Member
time taken: 0.160 seconds, Fetched: 3 row(s)
the table looks like this in hbase:
hbase(main):002:0> scan 'emp'
ROW COLUMN+CELL
123 column=cf1:name, timestamp=1422540225254, value=Ram
123 column=cf1:role, timestamp=1422540225254, value=TeamLead
456 column=cf1:name, timestamp=1422540225254, value=Silva
456 column=cf1:role, timestamp=1422540225254, value=Member
789 column=cf1:name, timestamp=1422540225254, value=Krishna
789 column=cf1:role, timestamp=1422540225254, value=Member
3 row(s) in 2.1230 seconds
Now I'm trying to update a value in this table
for exemple, I want to change the "role" of "Ram" from "Teamlead" to "Member",
which query should I use?

Answer 1

Assuming that you are trying to overwrite the previous value, from the hbase shell you can run the following:
put 'emp', 123, 'cf1:role', Member', 1422540225254
It's important to use the same timestamp as the previous entry if your goal is to overwrite.

Answer 2

You can do it with HIVE since v0.14+:
INSERT INTO TABLE hbase_table_emp VALUES (123, null, "Member");
You've got to provide the key you want to update and null values on fields you don't want to update... yeah, it's odd, like having to compile and run a MapReduce job for a single update, but it's also odd to pretend HIVE+HBase to work like a regular RDBMS and provide full ACID support in the process :)
For updating data I'd stick to the APIs provided by HBase (Stargate, Thrift, Native or even the hbase shell) and use HIVE just for mass imports and data analysis.