Database Reference
In-Depth Information
HBase is a columnar database, which means that instead of being organized
by rows and columns, it is organized by column families, which are sets of
related columns. Restructuring the data presented in
Figure 4.7
using the
columnar approach results in a layout that although similar is actually very
different (see
Figure 4.8
).
Figure 4.8
Columnar database structure
The columnar layout has many advantages over a relational model in the
context of handling big data, including the following:
• Can handle very large (even massive) quantities of data through a
process known as sharding
• Allows flexible data formats that can vary from row to row
• Typically scales linearly
For a more thorough discussion on columnar database capabilities and
HBase in general, check out the HBase website at
http://hbase.apache.org/
.
Defining and Populating an HBase Table
HBase is installed and configured for you as part of the Hortonworks Data
Platform.YoucanworkwithHBasedirectlyfromtheHBasecommandshell.
To define a table, you specify a table name and the column family or
families. In the following example, a basic customer table with a single
column family for addresses is created:
