In this section we will give an overview of the database schema and more specifically, we will focus on the SQLAlchemy specific schema.
In the following image you can see the tables of a fairly empty and new AiiDA database.
The database table names are marked at the top of every table (in deep green), the names of the the columns are below. The primary key of every table is highlighted with light green and the columns that participate to an index (apart from the index of the primary key) are highlighted with yellow color.
We will now briefly describe some of the database tables.
This table contains the information about all the Node instances stored in
AiiDA. Since the Node class has various sub-classes, there is a column (named
type) at the
db_dbnode table that indicates the type of the node.
Other columns are the
ctime and``mtime`` that correspond to the creation
time and the modification time of the node. There are foreign keys that link
this table with the
db_dbcomputer and the
It is also worth noting the
extras columns that are of
type JSONB. This type allows to store directly JSON snippets in such columns
while they can be indexed for fast searching.
The db_dbuser table contains user related information. It is worth noting that the stored password column contained hashed information.
db_dbgroup & db_dbgroup_dbnodes¶
The nodes can be grouped into groups. In the
db_dbgroup table contains
information about the group like the name, the type and the description of the
group. The nodes that are contained in the group are stored in
The calculations have a specific state which is needed for the user to verify their status but also for the daemon to submit them and/or fetch the results. This table contains the calculation id, the state that the calculation had and the timestamp of that state.
db_dbcomment table comments made by users to specific nodes are
db_dbworkflow table various information of the workflow are stored.
These are (among others) the name and the id of the workflow as well as various
The data related to each workflow is stored at the
This data could be the input or output of the workflow and can be stored as a JSON
value in database or in a text file.
The workflow is divided into multiple steps. Each step could be the new calculation
or the subworkplow. The general information like parent workflow id, status, name is
If the workflow step contains the calculation(s), the ids of the workflow step and the
calculation node are stored at
If the workflow step contains the subworkflow(s), the ids of the workflow step and the subworkflow
are stored at