Tiny MD-HBase

Tiny MD-HBase realizes a multi-dimensional data store on top of HBase, an 
ordered key-value store. The basic design is based on the MD-HBase paper[1][2].

Tiny MD-HBase is released as a toy version of MD-HBase. It focuses to 
provide simple working code samples of multi-dimensional data store features 
on HBase. Though it does not provide the complete features and highly tuned 
performance, I hope it helps you to understand how it works.


[1] Shoji Nishimura, Sudipto Das, Divyakant Agrawal, Amr El Abbadi,
    "MD-HBase: A Scalable Multi-dimensional Data. Infrastructure for Location Aware Services"
    In the 12th International Conference on Mobile Data Management (MDM) 2011, pp 7-16.
    http://www.cs.ucsb.edu/~sudipto/papers/md-hbase.pdf
[2] http://www.slideshare.net/nishimurashoji/mdm2011-mdh-base

********
License
*******

Tiny MD-HBase is released under Apache License 2.0.

***********
Compilation
***********

Just run 
> mvn package

**********
Deployment
**********

Put tiny-mdhase.jar under ${HBASE_HOME}/lib

**********
How to use
**********

Start HBase.
> cd ${HBASE_HOME}
> bin/start-hbase.sh


Put an entity at (x,y).
> bin/hbase tiny.mdhbase.Client put x y

Tiny MD-HBase only supports non-negative coordinates. Choose non-negative integers
for x and y. 'put' automatically generates entityID if you omit an ID.
If you want to specify entity ID, type
> bin/hbase tiny.mdhbase.Client put x y id

Now you put an entity at (x,y). Check entities at (x,y).
> bin/hbase tiny.mdhbase.Client get x y
[entityID, (x,y)]

'get' prints all entities at (x,y). In other word, 'get' is a point 
query that retrieves all entities at the query point.


If you put many entities into the store, the store may split a bucket.
Type the following command to inspect the index table.
> bin/hbase tiny.mdhbase.Client index

The default split threshold is 10. If you put more than 10 entities, you will
observe a bucket split. 

Let's try multi-dimensional range queries.
'count' prints the number of points within region (xmin, ymin)-(xmax, ymax).
> bin/hbase tiny.mdhbase.Client count xmin ymin xmax ymax


If you want to reset all entries, drop the table.
> bin/hbase tiny.mdhbase.Client drop