Hive Reader

Hive Reader plugin implements the ability to read data from Apache Hive database.

The main purpose of adding this plugin is to solve the problem of Kerberos authentication when using RDBMS Reader plugin to read Hive database. If your Hive database does not have Kerberos authentication enabled, you can directly use RDBMS Reader. If Kerberos authentication is enabled, you can use this plugin.

Example

We create the following table in Hive's test database and insert a record:

create table default.hive_reader
(
    col1 int,
    col2 string,
    col3 timestamp
)
stored as orc;


insert into hive_reader values(1, 'hello', current_timestamp()), (2, 'world', current_timestamp());

The following configuration reads this table to terminal:

{
  "job": {
    "setting": {
      "speed": {
        "byte": -1,
        "channel": 1
      },
      "errorLimit": {
        "record": 0,
        "percentage": 0
      }
    },
    "content": {
      "reader": {
        "name": "hivereader",
        "parameter": {
          "column": [
            "*"
          ],
          "username": "hive",
          "password": "",
          "connection": {
            "jdbcUrl": "jdbc:hive2://localhost:10000/default;principal=hive/_HOST@EXAMPLE.COM",
            "table": [
              "hive_reader"
            ]
          },
          "where": "logdate='20211013'",
          "haveKerberos": true,
          "kerberosKeytabFilePath": "/etc/security/keytabs/hive.headless.keytab",
          "kerberosPrincipal": "hive@EXAMPLE.COM"
        }
      },
      "writer": {
        "name": "streamwriter",
        "parameter": {
          "print": true
        }
      }
    }
  }
}

Save the above configuration file as job/hive2stream.json

Execute Collection Command

Execute the following command for data collection

bin/addax.sh job/hive2stream.json

Parameters

Configuration	Required	Type	Default Value	Description
jdbcUrl	Yes	list	None	JDBC connection information of target database
driver	No	string	None	Custom driver class name to solve compatibility issues, see description below
username	Yes	string	None	Username of data source
password	No	string	None	Password for specified username of data source, can be omitted if no password