Providing Parameters from a Database

The DatabaseParameterProvider at its core maps database rows to Parameters, specified by a Parameter Name Column and Parameter Value Column. The Parameter Group name must also be accounted for, and may be specified in different ways using the Parameter Grouping Strategy.

Before discussing the actual configuration, note that in some databases, the words 'PARAMETER', 'PARAMETERS', 'GROUP', and even 'VALUE' are reserved words. If you choose a column name that is a reserved word in the database you are using, make sure to quote it per the database documentation.

Also note that you should use the preferred table name and column name case for your database. For example, Postgres prefers lowercase table and column names, while Oracle prefers capitalized ones. Choosing the appropriate case can avoid unexpected issues in configuring your DatabaseParameterProvider.

The default configuration uses a fully column-based approach, with the Parameter Group Name also specified by columns in the same table. An example of a table using this configuration would be:

Table 1: Database table example with Grouping Strategy = Column
PARAMETER_CONTEXTS
PARAMETER_NAME	PARAMETER_VALUE	PARAMETER_GROUP
param.foo	value-foo	group_1
param.bar	value-bar	group_1
param.one	value-one	group_2
param.two	value-two	group_2

In order to use the data from this table, set the following Properties:

Parameter Grouping Strategy - Column
Table Name - PARAMETER_CONTEXTS
Parameter Name Column - PARAMETER_NAME
Parameter Value Column - PARAMETER_VALUE
Parameter Group Name Column - PARAMETER_GROUP

Once fetched, the parameters in this example will look like this:

Parameter Group group_1:

param.foo - value-foo
param.bar - value-bar

Parameter Group group_2:

param.one - value-one
param.two - value-two

Grouping Strategy

The default Grouping Strategy is by Column, which allows you to specify the parameter Group name explicitly in the Parameter Group Column. Note that if the value in this column is NULL, an exception will be thrown.

The other Grouping Strategy is by Table, which maps each table to a Parameter Group and sets the Parameter Group Name to the table name. In this Grouping Strategy, the Parameter Group Column is not used. An example configuration using this strategy would be:

Parameter Grouping Strategy - Table
Table Names - KAFKA, S3
Parameter Name Column - PARAMETER_NAME
Parameter Value Column - PARAMETER_VALUE

An example of some tables that may be used with this strategy:

Table 2: 'KAFKA' Database table example with Grouping Strategy = Table
KAFKA
PARAMETER_NAME	PARAMETER_VALUE
brokers	http://localhost:9092
topic	my-topic
password	my-password

Table 3: 'S3' Database table example with Grouping Strategy = Table
S3
PARAMETER_NAME	PARAMETER_VALUE
bucket	my-bucket
secret.access.key	my-key

Once fetched, the parameters in this example will look like this:

Parameter Group KAFKA:

brokers - http://localhost:9092
topic - my-topic
password - my-password

Parameter Group S3:

bucket - my-bucket
secret.access.key - my-key

Filtering rows

If you need to include only some rows in a table as parameters, you can use the 'SQL WHERE clause' property. An example of this is as follows:

Parameter Grouping Strategy - Table
Table Names - KAFKA, S3
Parameter Name Column - PARAMETER_NAME
Parameter Value Column - PARAMETER_VALUE
SQL WHERE clause - OTHER_COLUMN = 'my-parameters'

Here we are assuming there is another column, 'OTHER_COLUMN' in both the KAFKA and S3 tables. Only rows whose 'OTHER_COLUMN' value is 'my-parameters' will then be fetched from these tables.