Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Plain View
HBase, mail # user - Hbase Region Split not working with JAVA API


+
msmdhussain 2012-11-15, 08:42
+
Jean-Marc Spaggiari 2012-11-15, 18:43
+
Ameya Kantikar 2012-11-15, 18:55
+
msmdhussain 2012-11-15, 08:43
+
Anoop Sam John 2012-11-16, 03:40
+
msmdhussain 2012-11-16, 06:12
Copy link to this message
-
Re: Hbase Region Split not working with JAVA API
Stack 2012-11-16, 07:32
On Thu, Nov 15, 2012 at 10:12 PM, msmdhussain <[EMAIL PROTECTED]> wrote:
>
> i tried using the follwing command in hbase shell
>
>>> put
>>> 'Test','\x03\x00\x00\x00\x0A\x00\x00\x01:K\xF0\xC0@IN\x00\x00AS*******','C:input','25'
>
> but, row key length is 292 bytes its a composite key combination of
> time_seg+date+country_id+****+metrics_id (15 fields)
>
> it is stored in the correct region '\x03'
>
> and im converting all the above field values into  byte in java program and
> using the Put java api to insert the value to the hbase
>
> but, it get stored in the first region '\x01'
>
> i dont know why its happening when using java api
Your translation of region name must be incorrect.  It can be tricky
going from the String representation to the raw bytes.

Can you do something like figure out its ordinal position in .META.
then scan .META.  The rows in .META. are regionnames.  You can take
the Nth row in the .META. and use its bytes as the region name doing
your manipulations?

St.Ack