IH-ETL FAQ - IH-ETL Auto Recreate Temp Entity After Changing Import Set

[Siyou Li]

1. Delete a temp entity record in backend table (cmdb_inst_application_feed/sys_rte_eb_entity)

2. Go to IH-ETL mapping of above deleted entity, navigate to step 1 basic details

3. Pull a new import set or switch to a different import set

4. Go back to the entity backend table, check that the deleted entity is auto recreated again

[Siyou Li · Accepted solution]

What Are Temp Entities?

In IntegrationHub ETL, data flows through a three-layer metadata pipeline:

Raw Data -> Import (Imp) Entity → Temp Entity → Target (CMDB) Entity

The Temp Entity is an intermediate metadata layer automatically derived from the raw data schema. Each level/node in the incoming data (e.g., nested JSON arrays) maps to a corresponding temp entity. Temp entities act as the bridge that connects raw source fields to their CMDB class destinations.

When Does IH-ETL Auto-Recreate a Deleted Temp Entity?

The system will automatically recreate a missing temp entity only when all of the following conditions are met:

1. The integration uses a nested payload (i.e., the data source has "Data in single column" enabled — typically JSON/XML-based integrations).

2. The import set is changed — when the user loads new sample data or switches to a different import set.

3. The new import set is not empty — it contains at least one data row.

4. The raw data schema still contains the level/path that the deleted temp entity originally represented.

When these conditions are met, the system re-derives the required temp entity paths from the raw data schema and compares them against the existing metadata. Any path that exists in the data but has no corresponding temp entity (and no entity mapping) is automatically recreated along with its imp-to-temp entity mapping.

Note: If the import set ID has not changed (i.e., the same sample data is reloaded), or if the integration uses a flat payload, the system will not recreate deleted temp entities.

Why Is This Behavior Necessary?

The temp entity layer must mirror the structure of the incoming raw data for the ETL mapping to function correctly. Specifically:

• Each level in the raw data schema needs a corresponding temp entity. For nested payloads, each array-level node (e.g., items[*], items[*].components[*]) maps to its own temp entity. Source fields at that level are attached to that temp entity, and field mappings route data from those source fields to target CMDB class attributes.

• If a temp entity is missing but the raw data still contains fields at that level, those fields have nowhere to land. Without the temp entity, the system cannot create or maintain field mappings for that data level, and any data at that level will be silently dropped during transformation.

• When a new import set is loaded, the schema may differ from the previous one. The auto-recreation ensures that the metadata stays in sync with the actual data structure, so that all levels of the incoming data are properly represented and mappable — even if a temp entity was previously deleted (intentionally or accidentally).

In short: the temp entity is the structural anchor for field mappings at each data level. As long as the raw data still has that data level, then IH-ETL will recreate the temp entity for that raw data level input. Without it, the ETL pipeline cannot map fields at that level, and the integration will produce incomplete or incorrect CMDB data.