Creating HBase Indexes for Data Query

Scenario

HBase is a distributed database for key-value pairs. HIndex enables HBase to index based on the values of certain columns to narrow down the search scope and speed up searches.

Constraints

• Column families are separated by semicolons (;).
• Columns and data types must be contained in square brackets ([]).
• The column data type is specified by using -> after the column name.
• If the column data type is not specified, the default data type (string) is used.
• The number sign (#) is used to separate two index details.
• The following is an optional parameter:

  -Dscan.caching: number of cached rows when the data table is scanned.

  The default value is set to 1000.

• Indexes are created for a single region to repair damaged indexes.

  This function is not used to generate new indexes.

Creating an HBase HIndex

1. Install the HBase client. For details, see Using the HBase Client.
2. Go to the client installation directory, for example, /opt/client.

   cd /opt/client

3. Run the following command to configure environment variables:

   source bigdata_env

4. If the cluster is in security mode, run the following command to authenticate the user. In normal mode, user authentication is not required.

   kinit Component service user

5. Run the following command to access HIndex:

   hbase org.apache.hadoop.hbase.hindex.mapreduce.TableIndexer

   Table 1 Common HIndex commands

   Function	Command
   Add Index	TableIndexer-Dtablename.to.index=table1-Dindexspecs.to.add='IDX1=>cf1:[q1->datatype],[q2],[q3];cf2:[q1->datatype],[q2->datatype]#IDX2=>cf1:[q5]'
   Create Index	TableIndexer -Dtablename.to.index=table1 -Dindexnames.to.build='IDX1#IDX2'
   Delete Index	TableIndexer -Dtablename.to.index=table1 -Dindexnames.to.drop='IDX1#IDX2'
   Disable Index	TableIndexer -Dtablename.to.index=table1 -Dindexnames.to.disable='IDX1#IDX2'
   Add and Create Index	TableIndexer -Dtablename.to.index=table1 -Dindexspecs.to.add='IDX1=>cf1:[q1->datatype],[q2],[q3];cf2:[q1->datatype],[q2->datatype]#IDX2=>cf1:[q5]' -Dindexnames.to.build='IDX1'
   Create Index for a Single Region	TableIndexer -Dtablename.to.index=table1 -Dregion.to.index=regionEncodedName -Dindexnames.to.build='IDX1#IDX2'

   

   • IDX1: indicates the index name.
   • cf1: indicates the column family name.
   • q1: indicates the column name.
   • datatype: indicates the data type, including String, Integer, Double, Float, Long, Short, Byte and Char.