Skip to content

Latest commit

 

History

History
80 lines (56 loc) · 4.16 KB

no_sql_support.md

File metadata and controls

80 lines (56 loc) · 4.16 KB
Raw

NoSQL Support

You can implement your own custom implementation for NoSQL databases by

  • providing a StorageSystem which extends CustomStorageSystem,
  • providing a Parser which extends AbstractBeneratorDescriptorParser,
  • providing a Statement which extends Statement,
  • register the Parser in the BeneratorParseContext and
  • defining an own tag for the descriptor

The tag element for your own custom storage system will be parsed in the AbstractBeneratorDescriptorParser. It also will create a Statement, where the parsed tag elements will be handled and finally handed over to the CustomStorageSystem.

⚠ Information: For a custom storage system reference implementation see com.rapiddweller.platform.mongodb.MongoDBSystem

How to handle schema free storage systems

Pre parse generate tags

Some storage systems may not have a schema (e.g. MongoDB). But Benerator assumes that the storage systems knows and can provide a description about the data, which the storage system is managing. Therefore, we need to provide this description and to do so, the <pre-parse-generate> tag can be used. This tag pre parsed all <generate> tags and creates type descriptors for each <generate> and then provide them to the target custom storage system.

<pre-parse-generate target="db"/>

The target attribute will search for all <generate> tags where the consumer equals the target. If no data should be generated, but type descriptions still are needed, then you can set the count attribute in <generate> to 0. No data will be generated, but the storage system will still receive the type description.

⚠ Attention: The <pre-parse-generate> tag must be used after the storage systems are defined and before the first <generate>.

You can find an example of the usage in the demo scripts shop-mongodb.ben.xml and shop-postgres-mongodb.ben.xml.

Provide a meta-model

You can also provide a meta-model directly to a custom system storage by using the <meta-model> tag. The tag provides a subset of the element and attributes of the <generate> tag (only these which are important for the description of the data but not for the generation).

<meta-model type="db_category" consumer="mongodb">
    <id name="id" type="string" unique="true"/>
    <attribute name="name" type="string"/>
    <attribute name="parent_id" type="string"/>
</meta-model>

You can find an example of the usage in the demo scripts shop-mongodb.ben.xml and shop-postgres-mongodb.ben.xml.

MongoDB Support

Benerator supports MongoDB and can be used like this:

<mongodb id="mongodb" 
         host="{ftl:${mongoHost}}" 
         port="{ftl:${mongoPort}}" 
         database="{ftl:${mongoDatabase}}"
         user="{ftl:${mongoUser}}" 
         password="{ftl:${mongoPassword}}" 
         authenticationDatabase="{ftl:${mongoAuthenticationDatabase}}"
         authenticationMechanism="{ftl:${mongoAuthenticationMechanism}}"
         clean="{ftl:${mongoClean?c}}"
/>

The parameter authenticationDatabase and authenticationMechanism are optional. If authenticationDatabase is not set, the admin database will be used. If authenticationMechanism is not set, SCRAM-SHA-1 will be used.

An [environment file](./environment_files.md#Environment Files) can also be used:

<mongodb id="db" environment="mongodb" system="target"/>

The parameter clean is optional and if it set to true, the database will be cleaned at start (all data removed). The default value is false. Be careful with this parameter

You can find an example of the usage in the demo scripts shop-mongodb.ben.xml and shop-postgres-mongodb.ben.xml.

Limitations

  • The import of nested entities is currently not possible.
  • The import of entities requires type descriptions.
  • Types cannot be automatically derived. Please provide a type in <generate>. Otherwise, the default type is string.
  • Queries/Selectors uses comamnd from db.runCommands(command) internal. The commands find and aggregate are implemented. The commands count, distinct and mapReduce could also be used, but are not tested yet.