How to rename fragements Without losing historical data

Forum
1

Raising this on behalf of a platform user

I am managing an IoT implementation for monitoring steam and water consumption at our industrial sites. We use Modbus TCP with PSSystec devices to connect our industrial metering equipment to the Cumulocity IoT platform.

Issue:

When I change a datapoint name in the protocol, the system creates a new datapoint instead of updating the existing one. This creates two problems:

  1. Historical data remains tied to the old measurement type
  2. I end up with duplicate measurements (the original and the renamed version)

In our industrial environment, measurement value types are often assigned temporarily during commissioning. The correct KKS (power station identification system) designation may come later or need correction. Currently, when we rename a measurement, all previous data is effectively split between two measurement types.

My Question:

Is there a way to rename existing measurement types while preserving the historical data? Alternatively, can we migrate historical data from one measurement type to another?

2

Initial response i gathered from Cumulocity Team

Measurements in Cumulocity IoT are designed to be immutable (cannot be modified after creation). This is why changing a datapoint name creates a new measurement type rather than updating the existing one.

To rename measurement types while preserving historical data, you’ll need to follow a migration process:

  1. Read existing measurement series from the old fragment/series name
  2. Create new measurements in the desired format/name
  3. Delete the old measurements if desired (optional, but recommended to avoid confusion)

Option 1: Custom Script You can create a custom script to perform this migration.

Option 2: Go-C8Y-CLI Tool You can also use command-line tool (go-c8y-cli) to do the same . The documentation includes examples for copying measurements from one device to another. While renaming measurements is slightly more complex, you can adapt these examples for your needs.

1 Like
3

One thing to add:

  1. If the data used in the fragment or series name of the measurmeent is dynamic, it shouldn’t been used there (if possible). Mainly static information should be used for it e.g. having a key that identifies for what the measurement stands for. In addition the dynamic data can be either set on the device (the measurment is created for) or in the measurement itself in a custom fragment. So in your given example KKS → Device, Measurement → Water consumption data points.
    A good practices could be to write the measurements on a generic device (containing a fragment with dynamic data which could be changed) so with full history across all dynamic assignment and in addition to a dynamic device (if required).