3.6 Bulk Creation of Elements and Attributes
Bulk creation of elements and attributes is a practical concern for many users. Real-world scenarios involve thousands or even tens of thousands of elements and attributes — it is simply not feasible to create them one by one by hand.
TDengine IDMP has multiple bulk modeling capabilities built in. It can automatically generate element templates, element instances, and attributes from a TDengine TSDB schema, a CSV file, or an OPC structure — so that a trial environment has a working model "the moment it is installed", and a production deployment is compressed from "weeks of manual configuration" to "minutes of automatic construction".
3.6.1 Scenarios Suitable for Bulk Modeling
If you are in any of the following situations, you can start with bulk modeling to build the skeleton of the data model, then refine and adjust it locally:
- You are already storing time-series data in TDengine TSDB — you can reverse-generate the model directly from the existing supertables, child tables, and tag structure;
- You have a tag system organized under some naming convention (paths like
Plant.Line1.Machine3) that can be mapped directly into an asset tree; - The number of assets and supertables is large, and you want to centrally manage the mapping rules in a CSV / spreadsheet;
- The data comes from an OPC server and has been ingested into TDengine TSDB via OPC-UA / OPC-DA, and you want to preserve the original OPC node structure when modeling.
3.6.2 Bulk Modeling Methods in IDMP
On the TDengine connection details page under Admin Console → Connections → [connection name], IDMP provides the following four independent bulk-modeling methods that can be used as needed:
| Method | Suitable Scenario |
|---|---|
| Simple Import | The asset hierarchy is already encoded in a supertable tag (e.g. location = Plant.Line1.Machine3), and you want a one-click generation of the full element template + element + attribute set |
| Map Supertables to Elements | Tags don't carry a hierarchy, or the model is "one supertable per measurement", and multiple supertables need to be combined under the same element template |
| Import from CSV | Centrally configure a large number of supertable / attribute mappings through a spreadsheet — well suited for bulk go-live and team collaboration |
| Import from OPC | Reverse-build the asset model from OPC-structured data already stored in TDengine TSDB, preserving the original OPC node path structure |
All four methods share these common characteristics:
- Auto-creates templates + elements + attributes: a single configuration generates element templates, element instances, and attributes, and binds the attributes to the correct TDengine metric columns or tags;
- Automatic synchronization: after the import task runs, IDMP keeps watching TSDB metadata changes, and new child tables added to a configured supertable are automatically synced as new elements, with no manual intervention;
- Repeatable configuration: supports Rebuild and remapping, making it easy to keep tuning the model during iteration.
3.6.3 Bulk Creation via Copy and Paste
In addition to the external-data-source import methods above, most objects inside IDMP — including elements, attributes, panels, and analyses — support copy and paste operations. This is a very practical "lightweight bulk modeling" technique, well suited for extending and reusing an existing model.
- Flexible element granularity: copy and paste is supported not only on a single element, but also on a mid- or upper-level node containing multi-level child nodes in the element tree. The subtree structure, attributes, and attribute bindings are all copied together;
- Works across object types: attributes, panels, and analyses likewise support copy and paste, making it easy to migrate configurations quickly between different elements or templates;
- Asynchronous execution for large batches: when a copy-and-paste operation involves a large number of objects (such as copying a subtree containing hundreds or thousands of child nodes), the system automatically switches to an asynchronous mode and runs in the background. The front-end user does not have to wait and can keep working on other tasks; the results are refreshed once the job completes.
Copy and paste can be combined with the import methods above: first use import to quickly generate the model skeleton, then use copy and paste to extend, reuse, and fine-tune local structures.
3.6.4 Refining the Data Model
After bulk modeling, users can continue to refine and locally adjust the data model:
- Fill in units, high / low limits, target values, categories, descriptions, and other contextual information for elements and attributes (see 3.3 Data Contextualization);
- Establish the necessary element references (see 3.5 Relationships and Industrial Ontology) — for example, weakly referencing a shared measurement point under multiple process stages;
- Define analyses, panels, dashboards, and notification rules uniformly at the element-template level, so that every asset of the same kind gets standardized visualization and monitoring in one shot.
For the complete bulk-modeling reference — including the field descriptions, expression syntax, typical examples, and advanced usage of each method — see 12.3 Building Data Models from TDengine TSDB.
