MindsDB Schema
Initially, MindsDB comprises three system databases and one default project, as follows:
information_schema
stores metadata of all the objects such as handlers, databases, AI engines, models, jobs, and more.log
stores log data of models and jobs.files
, which is initially empty, stores all files uploaded to MindsDB.mindsdb
is the default project for storing models, views, jobs, triggers, and agents.
List all databases by running the following SQL commands:
Here is the output:
The information_schema
Database
The information_schema
database contains all the system tables that correspond to MindsDB objects as follows:
Name | Description |
---|---|
HANDLERS | Stores all (data and AI) handlers, which are data integrations and AI integrations supported by MindsDB. |
DATABASES | Stores all data sources connected to MindsDB. Note that corresponding handlers are required, and you can connect only the data sources supported by MindsDB after installing the required handler dependencies. |
ML_ENGINES | Stores all AI/ML engines configured at MindsDB. Note that corresponding handlers are required, and you can connect only the AI/ML engines supported by MindsDB after installing the required handler dependencies. |
MODELS | Stores all models deployed within the MindsDB ecosystem. Note that you can create and deploy a model only after configuring the corresponding AI/ML engine. |
VIEWS | Stores all views created in MindsDB. |
JOBS | Stores all jobs that facilitate workflow automation. |
TRIGGERS | Stores all triggers that facilitate workflow automation. |
AGENTS | Stores all AI agents created in MindsDB. |
SKILLS | Stores all skills that can be assigned to AI agents. |
KNOWLEDGE_BASES | Stores all knowledge bases that can be assigned to AI agents as skills. |
CHATBOTS | Stores all chatbots that comprise an AI agent and a chat interface. |
Some of the objects, including DATABASES
, ML_ENGINES
, and MODELS
, may contain sensitive information in the form of API keys or passwords. MindsDB hides this sensitive information by default.
If you want to expose this sensitive information in the output when querying these objects, set the show_secrets
flag to true
.
And to hide them back, set it to false
.
Use the SHOW
command to list all objects as follows:
For instance, list all OpenAI models from the mindsdb
project that contain ai
in its name.
Another example of how to query for the available data and AI handlers:
Before you can connect a data source using a data handler or create a model using an AI handler, make sure that the IMPORT_SUCCESS
column reads true. If it reads false, then install the dependencies for this handler before using it.
The mindsdb
Project
MindsDB enables you to group all objects within projects. Projects can store models, views, jobs, triggers, agents, skills, knowledge bases, and chatbots.
Projects store all objects except for handlers, connected data sources, and configured AI/ML engines.
Note that based on the available handlers, you can connect a data source to MindsDB or configure an AI/ML engine within MindsDB. Having done that, you can, for instance, create a view with data from the connected data source and store it inside the project, or create a model based on the configured AI/ML engine and store it inside the project.
MindsDB provides the default mindsdb
project where all objects created without defining a project are stored.
Learn more about how to create and manage projects here.
The files
Database
It is another default database that stores all the files uploaded to MindsDB.
Here is how you can upload files to MindsDB.
Was this page helpful?