> ## Documentation Index
> Fetch the complete documentation index at: https://bloodhound.specterops.io/llms.txt
> Use this file to discover all available pages before exploring further.

# OpenGraph FAQ

> The following are common questions about OpenGraph

![Applies to BloodHound Enterprise and CE](https://mintcdn.com/specterops/tTIczgde9H07oLXf/assets/enterprise-AND-community-edition-pill-tag.svg?fit=max\&auto=format\&n=tTIczgde9H07oLXf\&q=85\&s=ad49a576589f4d2a8081df77d07fdf56)

> [accordion] Can I define my own icons in OpenGraph?
>
> Yes you can!
> You can find all the details [here](https://bloodhound.specterops.io/opengraph/developer/custom-icons.md)

> [accordion] Can I delete data ingested via OpenGraph?
>
> Yes you can!
>
> You can remove generic data by using one of the following three options:
>
> 1. Cypher commands in the Explore UI
>
>    Requires `enable_cypher_mutations: true` in config. For more info [click here](https://bloodhound.specterops.io/manage-bloodhound/bh-config.md#enable-cypher-mutations)
>
> 2. Admin UI → Database Management page
>
>    The following checkboxes will be displayed:
>
>    * All Data
>
>    * Active Directory Data
>      * HasSession edges
>
>        Use the **HasSession edges** option when you need to refresh time-sensitive session data without deleting the rest of the collected graph.
>
>        > [!NOTE]
>        >
>        > Manual **HasSession** edge deletion is generally only necessary in BloodHound Community. In BloodHound Enterprise, **HasSession** edges reconcile automatically by [retention](https://bloodhound.specterops.io/collect-data/enterprise-collection/data-retention.md#data-retention) and only need manual deletion when you want an immediate session data refresh.
>
>    * Azure Data
>
>    * … X Data
>
>      There will be 0 or more checkboxes here that allow you to delete any data that had been ingested with a `source_kind` provided.
>
>    * Sourceless Data
>
>      Checking this box will delete all entities that do not have a kind that can be found in the `source_kinds` table.  You can observe the `source_kinds` your BloodHound instance is currently aware of by calling `GET /api/v2/graphs/source_kinds`.
>
> 3. API: [`/api/v2/clear-database`](https://bloodhound.specterops.io/reference/database/delete-your-bloodhound-data)

> [accordion] Do custom nodes & edges work with Search and Pathfinding?
>
> Yes! BloodHound supports **Search** and **Pathfinding** for [*structured*](https://bloodhound.specterops.io/opengraph/extensions/manage.md#structured-graphs) graphs.
>
> For [*generic*](https://bloodhound.specterops.io/opengraph/extensions/manage.md#generic-graphs) graphs, only **Search** is supported.
>
> Pathfinding, findings, and analysis require traversable relationship kinds defined in an extension definition schema. Generic graph data does not define traversability and is treated as non-traversable.
>
> > [!NOTE]
> >
> > All forms of OpenGraph data support Cypher querying in general.

> [accordion] Why do nodes disappear when using Search or Pathfinding?
>
> This usually happens when a node's object ID contains a colon (`:`).
>
> Colons are not currently supported in object IDs in BloodHound.
>
> * You cannot enter an object ID that contains a colon in the **Search** bar.
> * You can still select a node by name, even when its object ID contains a colon.
> * If you select a node by name, **Search** or **Pathfinding** can complete, but the node disappears from the search bar afterward.
>
> To avoid this issue, use OpenGraph node IDs that do not contain colons.

> [accordion] How can I add my project to the OpenGraph library?
>
> Have you built a cool project using OpenGraph and want it featured here? Already got your project in the list and need to update something? Open a ["Library Change" issue](https://github.com/SpecterOps/bloodhound-docs/issues) on the BloodHound Docs repo and we'll get it added for you!
>
> Submissions from the community will have a  icon next to them while those by SpecterOps employees will have a SpecterOps icon.

> [accordion] Why is it taking so long to ingest OpenGraph data?
>
> Slow ingestion speeds are commonly seen when using a Neo4j graph database and are typically resolved by switching to a PostgreSQL backend. For full, officially supported OpenGraph functionality, use a PostgreSQL backend (see [Requirements](https://bloodhound.specterops.io/opengraph/best-practices.md#requirements)).
>
> To switch to a PostgreSQL graph database, set `"graph_driver": "pg"` in your `bloodhound.config.json` file and restart the BloodHound services/containers. You will need to re-ingest any data into the new database.

> [accordion] Why do AD/AZ nodes disappear after I delete OpenGraph data?
>
> The initial OpenGraph implementation introduced the `source_kind` metadata field to let you create a source and quickly apply it to all nodes in the payload. This preliminary process applies the `source_kind` to all nodes in the payload—whether you define or reference them—which can create unintended side-effects when building hybrid paths that connect OpenGraph nodes to Active Directory (AD) or Azure (AZ) nodes.
>
> If you reference AD/AZ nodes when defining an edge in an OpenGraph payload that uses this metadata field, the system inadvertently applies the `source_kind` to those non-OpenGraph nodes. If you delete this source and its OpenGraph data, then those AD/AZ nodes are also deleted.
>
> > [!NOTE]
> >
> > A more comprehensive solution is currently being designed to prevent this issue. The process described below is a workaround until the new design is implemented.
>
> **What this means:**
>
> When you delete OpenGraph data that references AD/AZ nodes with an OpenGraph `source_kind`, those AD/AZ nodes may also be deleted even though they are part of the core AD/AZ graph.
>
> **Recommended workaround:**
>
> Use a two-step upload process when building hybrid paths.
>
> * **First payload (isolated subgraph)**: Upload only your OpenGraph nodes and edges and set `source_kind` in `metadata`. Ensure that every node has at least one edge within this payload (for example, connect nodes to each other or to a root node and add a containment edge, such as `(:OGRoot)-[:OGContains]->(:YourNode)`). Do not link to AD/AZ objects in this payload.
>
> * **Second payload (linking AD/AZ nodes)**: Link the isolated subgraph's nodes to the AD/AZ nodes that connect your subgraph to existing AD/AZ entities. This avoids cross-source side effects when `source_kind` is present.
>
>   > [!NOTE]
>   >
>   > Do **not** define `source_kind` in the second payload.
