All versions of this manual

Importing graph data: Updating your graph data

When saving a visualization or a case, Linkurious Enterprise only stores the ID references to the nodes and edges stored in your graph database.
When updating your graph data you need to make sure your graph engine preserves those references to avoid any data loss.

There are 2 different strategies for updating your graph data. Each strategy has its own advantages and risks.

1. Updating your graph data incrementally.

This strategy consists in adding, removing or updating the nodes and edges within an existing database.
You can use this strategy when you know exactly the nodes and edges that have changed since your last database update.

After performing this update the following changes will be reflected in Linkurious Enterprise:

  • The nodes and edges modified will be reflected in your visualizations when you reopen them.
  • The nodes and edges added or modified will be properly searchable automatically or after a manual re-indexing operation depending on your search strategy. Learn more.
  • The nodes and edges removed from your graph database will be removed from your visualizations when you reopen them.

Neo4j and ID recycling

When you remove nodes and edges from Neo4j, the IDs of the deleted nodes and edges are recycled when you later create new nodes and edges. The consequence is that your existing visualizations in Linkurious Enterprise may contain references to new nodes and edges unrelated to the existing context.
If you plan to incrementally remove data from Neo4j, we recommend you use alternative identifiers before creating your visualizations.

2. Rebuilding your graph database.

This strategy consists in recreating your graph database from an external source.
You consider the current graph database stale, and you create a new graph database with fresh data.
This strategy is useful when your current infrastructure does not allow you to keep track of which nodes and edges are updated.

When you rebuild your graph database, new IDs will be assigned to all your nodes and edges, consequently breaking the references stored by Linkurious Enterprise. If you are planning to use this strategy, it is very important you configure alternative identifiers before creating your visualizations (only available with Neo4j).

In the unfortunate scenario that you have performed a database rebuild without configuring the alternative identifiers, Linkurious Enterprise will prevent access to your visualizations in an effort to prevent any data loss.

As an administrator, you can perform the following actions:

  1. Restore the previous version of the graph database from a backup
  2. Configure alternative identifiers (only available with Neo4j)
    1. After performing this configuration, you will need to re-open and save each visualization and case individually in order to apply the new configuration.