How to set pre-partitions for an HBase table - ApsaraDB for HBase

The pre-partitioning feature of ApsaraDB for HBase helps you design effective rowkeys to prevent data hot spots. This topic describes how to set pre-partitions for an HBase table.

Configuration method

To set pre-partitions for an HBase table named test, use the following example statement. The table below describes the fields used in the example.

create'test',{NAME => 'f1',COMPRESSION => 'snappy' }, { NUMREGIONS => 50, SPLITALGO => 'HexStringSplit' }

Field	Description
NAME	The column family name of the HBase table.
COMPRESSION	The data compression format for the HBase table. For more information, see Data compression and encoding.
NUMREGIONS	The number of regions. Calculate the number of regions based on a storage size of 6 GB to 8 GB per region. For large clusters, increase the number of regions.
SPLITALGO	The algorithm to split Rowkeys. ApsaraDB for HBase provides three split algorithms. The scenarios for each algorithm are described below: HexStringSplit: Use for Rowkeys that have a hexadecimal string as a prefix. DecimalStringSplit: Use for Rowkeys that have a decimal number string as a prefix. UniformSplit: Use for Rowkeys that have a completely random prefix. For more information about how to design Rowkeys, see How to design a Rowkey.