Specifying MapR-DB Binary Tables When Using the C APIs
In APIs in which you must pass the name of a MapR-DB table, you can specify the name in one of two ways:
- Specify the complete path to the table in this format:
/mapr/<cluster>/<volume>/<table>
For example, if you were to use the nhb_admin_table_add_column_family
API to add a new column family to a table, the name of the cluster wasnewyork
, the name of the volume wasvol1
, and the name of the table wastable1
, you would specify the table with this path:/mapr/newyork/vol1/table1
- Specify the name only. This method requires that you first modify the
hbase.table.namespace.mappings
property to thecore-site.xml
file. For Hadoop 1, edit the file/opt/mapr/hadoop/hadoop-0.20.2/conf/core-site.xml
. For Hadoop 2, edit the file/opt/mapr/hadoop/hadoop-2.4.1/etc/hadoop/core-site.xml
. Here is an example:
<property>
<name>hbase.table.namespace.mappings</name>
<value>*:/tables_dir1</value>
</property>
In this example, all tables that you create with the C APIs and all tables that you access
will be in the directory /tables_dir1
. If you specified the table name
table1
in a C API, the full path to the table would be
/tables_dir1/table1
.