aggregate=sum for synthetic timeseries

Hi!

It is possible today to sum using the Synthetic Time Series API, please take a look at the developer guide:

https://developer.cognite.com/dev/concepts/resource_types/synthetic_timeseries.html

Does this satisfy what you need?  If not, please can you give more details about what you’d like to do so we can better understand the request.

Kind Regards, 

Glen

Hellо Glen,

Based on what I see in the documentation, the aggregate parameter must be one of the following values: interpolation, stepinterpolation, or average. If I pass aggregate=sum, then the CDF returns the following error:

"Please check your syntax. Error: sum is not one of the valid aggregates: AVERAGE, INTERPOLATION, STEP_INTERPOLATION."

Is it possible to obtain the sum without using aggregation?

Regards,

Ievgen

From the Functions section higher in the document it describes using mathematical operators to combine two time series into a new synthetic time series, for example:

combine time series: TS{id=123} + TS{externalId='my_external_id'}

Is this what you’re trying to do?

 

Hi,

I mean an equivalent of the sum aggregation that we already have for regular time series.

https://developer.cognite.com/dev/concepts/aggregation/#sum

Regards,

Ievgen

AH! OK, now I think I understand.

So you want to sum the values that are generated within a time range, within a synthetic time series?

Right, like on this picture from the documentation

Yep, I understand now.  Sorry about that and yes we don’t have that function in Synthetic TS today.

I believe you may be able to accomplish the same result however using Charts calculations.  @Arun Arunachalam is the product manager for that, hopefully he can help whilst we consider the feature request for STS.

 

Hi,

Is there any update on when this feature could be available? This would greatly simplify our data analysis process.

Hi ​@Yatish Pitta,

I’ve taken over the Time Series and OT product management role from Glen.

Unfortunately, we do not currently have this capability on our roadmap, and as of right now we do not have the capacity to add it in 2025. We will revisit this for the 2026 timeframe.

​@Everton Colling  Are there any plans to implement this?

Hi ​@Aditya Kotiyal!

No, we do not have this extension planned in the roadmap, and we don’t have capacity to bring it in the first half of 2026. In any case, before we consider it, it would be valuable to understand how this would be helpful. Considering we already support “sum” aggregate in regular datapoints requests, what use case would be simplified/enabled by exposing “sum” in synthetic timeseries requests?