Create an ApsaraDB for HBase dimension table

What is ApsaraDB for HBase?

ApsaraDB for HBase is a cost-effective cloud-based intelligent NoSQL service that provides high scalability and is compatible with open source HBase. ApsaraDB for HBase provides benefits such as low storage costs, high throughput and scalability, and intelligent data processing. ApsaraDB for HBase supports many core services of Alibaba such as Taobao recommendations, risk control for Ant Credit Pay, advertising, data dashboards, Cainiao logistics track, Alipay transaction records, and Taobao Mobile messages. ApsaraDB for HBase is a fully managed service that provides enterprise-level capabilities such as the processing of petabytes of data and tens of millions of concurrent requests, quick scaling within seconds, low response latency within milliseconds, high availability across data centers, and global distribution.

Prerequisites

• An ApsaraDB for HBase cluster is purchased and an ApsaraDB for HBase table is created. For more information, see Purchase an instance.
• A whitelist is configured for the ApsaraDB for HBase cluster. For more information, see Configure a whitelist.

Limits

Only Flink that uses Ververica Runtime (VVR) 2.0.0 or later versions supports ApsaraDB for HBase connectors.

DDL syntax

CREATE TABLE hbase_dim(
  rowkey INT,
  family1 ROW<q1 INT>,
  family2 ROW<q2 STRING, q3 BIGINT>,
  family3 ROW<q4 DOUBLE, q5 BOOLEAN, q6 STRING>
) with (
  'connector'='cloudhbase',
  'table-name'='<yourTableName>',
  'zookeeper.quorum'='<yourZookeeperQuorum>'
);

• Column families of an ApsaraDB for HBase table must be declared as the ROW type. Each column family name is the field name of a row. In the DDL syntax, the following column families are declared: family1, family2, and family3.
• A column in a column family corresponds to a field in a row. The column name is the field name. In the DDL syntax, the q2 and q3 columns are declared in the family2 column family.
• In addition to the fields of the ROW type, only fields of the atomic type such as STRING and BIGINT can exist in an ApsaraDB for HBase table. The fields of the atomic type are considered as row keys of the table, such as rowkey in the DDL statement.
• The row key of an ApsaraDB for HBase table must be defined as the primary key of the result table. If no primary key is defined, the row key is used as the primary key.

Parameters in the WITH clause

Parameters in the CACHE clause

Data type conversion

Sample code

CREATE TABLE datagen_source (
  a INT,
  b BIGINT,
  c STRING,
  `proc_time` AS PROCTIME()
) with (
  'connector'='datagen'
);

CREATE TABLE hbase_dim (
  rowkey INT,
  family1 ROW<col1 INT>,
  family2 ROW<col1 STRING, col2 BIGINT>,
  family3 ROW<col1 DOUBLE, col2 BOOLEAN, col3 STRING>
) WITH (
  'connector' = 'cloudhbase',
  'table-name' = '<yourTableName>',
  'zookeeper.quorum' = '<yourZookeeperQuorum>'
);

CREATE TABLE blackhole_sink(
  a INT,
  f1c1 INT,
  f3c3 STRING
) with (
  'connector' = 'blackhole' 
);
  
INSERT INTO blackhole_sink
     SELECT a, family1.col1 as f1c1,  family3.col3 as f3c3 FROM datagen_source
JOIN hbase_dim FOR SYSTEM_TIME AS OF src.`proc_time` as h ON src.a = h.rowkey;