Product User Community

As Cognite provides APIs to access RAW tables from database, is there any way to get the actual datatype (text/bool/timestamp etc.) of the table’s column.

Hi community!Monitoring asset health, performing root cause analysis, and doing on-the-fly troubleshooting can be a complex process for both seasoned experts and new engineers. In this video I'll be sharing some tips on how to gather and analyze data efficiently, which we believe will boost your confidence in managing asset health.I'd love to hear your thoughts and encourage you to share your experiences and insights to help improve our asset health program.Cheers,Andreea.  

Have you already tried and used the latest features released in our Industrial Canvas? We would like to hear your feedback! 🙌Here is the full list of features and how you can use them:Managing your canvasCanvas visibility can now be set to either private or public. By default, when you create a canvas, it will be “marked” as private. You can change it to public by using the Share option while in the canvas. View live datapoints signalled with traffic lights based on basic rulesPin time series connected to assets in engineering diagrams and visualize their “live” values (the latest value is refreshed every 30s). Add simple rules on these values. For example “if the temperature is higher than 30°C, show the tag in colour red”.Leave comments and tag colleagues in the canvas You can easily add comments while working in the canvas and tag your colleagues. You can find all the comments in the “Comments panel” where you can also search for the content.  General fixes, tweaks and small improv

Hi I received a few P&ID diagram where the tags aren’t spelled out in-place (because the space perhaps was limiting), but just an indexed number is given. And in a separate table, the index is mapped with the tag name.Please see that attached image which can provide further illustration. I am not sure if the OCR + contextualization logic can map the index out and place the tag at the correct place. I am thinking it is unlikely. In such case, can you please provide clues on how one can map it - some amount of manual steps and tweaks, if expected, are fine. Thanks for your inputs! 

Hi Team, We are working on the Cognite Pi Extractor to fetch the data from OSIsoft PI Data Archive. From the documentation we understood that we can use the OSIsoft PI Web API or PI SDK to retrieve the data we need from OSIsoft PI Server. But it is not specifically mentioned in Cognite PI Extractor's documentation that which method it is internally using. Is it using PI SDK or PI AF SDK or PI Web API?

Hi team,I have a Honeywell PHD source system. I have a got a suggestion to use Cognite OPC Classic Extractor. But I can see Extractor suggested is beta version. Can I know, it is efficient to use the Cognite OPC Classic Extractor, or we should go with custom extractor.If we use ,Cognite OPC Classic Extractor it connects over COM+ or DCOM to OPC servers and streams live and historical data into CDF. Can you please elaborate on COM+COM+ or DCOM connectivity to OPC Servers. How to connect with Honeywell PHD server using either of the ways?    Thanks

Hello,I’m trying to publish a new version of an existing data model and I’m receiving this error message:containers must be unique across space and externalId, regardless of caseCan you please help me with some details regarding where to look here?Thanks,Regards,

I am trying to establish Parent-Child relationships between two types of events ingested and transformed in CDF which are Workorders and Maintenance Notifications. I have two different datasets for each where the Raw data is staged and have also given corresponding labels (Work_order, Notification) for dataset and later these workorders and Maintenance notifications are linked to Assets in CDF using transformations. Now I need to establish the relationship between the workorders and Maintenance notifications. Although I went through the Documentation for relationship ( Relationships | Cognite Documentation) not able to get the hold of it. Any reference SQL for how relationships are established would also do.

There is a problem in fetching raw tables records through SDK. This was working till last week and something has changed now.  client.raw.rows.retrieve_dataframe(db_name="Eashwar_MOTDB-db",table_name="lp_input",limit=None)Now getting this error below. Please advise.--------------------------------------------------------------------------MissingSchema Traceback (most recent call last)Cell In[6], line 1----> 1 client.raw.rows.retrieve_dataframe(db_name="Eashwar_MOTDB-db",table_name="lp_input",limit=None,columns=None)File /lib/python3.11/site-packages/cognite/client/_api/raw.py:538, in RawRowsAPI.retrieve_dataframe(self, db_name, table_name, min_last_updated_time, max_last_updated_time, columns, limit) 514 """`Retrieve rows in a table as a pandas dataframe. <https://developer.cognite.com/api#tag/Raw/operation/getRows>`_ 515 516 Rowkeys are used as the index. (...) 535 >>> df = c.raw.rows.retrieve_dataframe("db1", "t1",

We are currently exploring potential use cases with our team, and one of them involves implementing data quality techniques. I was wondering if anyone could share their experience in order to identify the best approach within the Cognite Platform for conducting data quality checks to ensure that all fields in critical tables are populated correctly.In the past, we have implemented a similar solution with a different vendor, utilizing a system that can prevent the flow of erroneous data through our data pipelines by employing validation and integrity checks based on declarative quality expectations.Thank you in advance.

List Projects API (https://{cluster}.cognitedata.com/api/v1/projects) - no corresponding way to use this from Python SDK.Is there any specific reason why this method to get the CDF Projects is not available in python sdk.If it is present I was not able to find it in the documentation. https://cognite-sdk-python.readthedocs-hosted.com/en/latest

We want to extract data from Bentley AssetWise APM. Can Cognite OPC UA Extractor be used for this use case? If yes, how? If no, which Extractor from Cognite’s prebuilt extractors is suitable for this?

In the below example, suppose “Heating cable” needs to be deleted. How can this be achieved? If “Heating cable” needs to be renamed like “Heating-Cable”, how can this be achieved? If “Heating Cable” needs to move from Pump_A to Pump_B, how can this be achieved? SYSTEM 11└───Pump (Pump_A)│   └───Heating cable└───Pump (Pump_B)

I am trying to connect to an MQTT broker with Cognite’s MQTT Extractor.I know that the broker is fully functional, verified by several tools.However, when completing the “Add MQTT Topic”.Error code indicates that the API does not accept “Datapoints” as “Target type”, even though “Target type” is a mandatory field in the interface window. And: “Datapoints” is the only available selection.Any quick feedback to progress my testing is appreaciated! 

Even after I change the name of a synthetic tag with the new InField 2.0 feature release, I still only see the original tag name in the first edit view. I have to click on the pencil icon to edit the tag, and then I will be able to see what the new name I changed it to, is. The name that shows up in the first edit task view should show as the updated tag, rather than the original tag name.

We are having a multilevel asset hierarchy in cdf, and ground level assets are only having timeseries linked.we are trying to create Relationships between every parent asset and its child asset’s Timeseries.we used python sdk code snippets to do this. Which are been attached to this for your reference.In python sdk, after running the python relationship code, it is showing relationships are established. But in UI it is not showing from Assets, But we are able to see under Timeseries.even after we tried interchanging the Source and target between assets and timeseries, same result.one unique observation is that we could see the relationship established just for 1 asset which is having 3 timeseries linkage. rest others are not showing(they have more than 20 relations, even 500 for root asset.)we would like to know are we following correct manner, or any restrictions are from CDF tool. plz help us in this regard.from cognite.client import CogniteClientfrom cognite.client.data_classes impo

Hello I am trying to run entity matching on the data that is in FDM (Data modelling) and not in the classic assets model. I don’t see an option to fetch such items in the entity matching screen (contextualize menu). Is this by design, do we expect a support to search and select such custom models for matching in the future. Also a similar case for data explorer, even when I have updated all my records to have a data set id, I am not able to search for them in the Data explorer. Just wondering when the explorer can start supporting such data. PS:I imagine, the entity matching can also work if I invoke the APIs directly,  but then I’ll have to extract names + description myself and also need to link the results back to their original ids. A visual ways in CDF Fusion UI will be very helpful. RegardsAlex 

We are pleased to announce an update to the Search, Filtering and Aggregations documentation, detailing enhanced search, filtering and aggregation capabilities for our Assets, Documents, Events, Sequences and Time Series APIs.These additional features are already implemented in the April 2023 update to Data Explorer and are available to use natively within the service APIs. The documentation provides guidance on how to implement advanced filtering, sorting, search and aggregation functionality (which is available on both the standard fields and metadata), so that application users can use much more powerful and expressive search queries to find exactly the data they are looking for.Use is recommended for existing, mature projects, and apply to use cases for the Assets, Documents, Events, Sequences and Time Series APIs.  For new projects and use cases, customers are recommended to work with the newer Data Modelling APIs.The updated documentation may be found here Tell us what you think

Hello All,In the past, we encountered some performance issues when extracting data from the PI System using PI API. As an alternative, we have been using PI SDK to extract data from both PI Archive and AF. The question is, does Cognite PI Extract utilize PI SDK, and are there any performance limitations?

Hi,We saw quite frequently an error message like below:"could not add instances: 500 - Internal server error. Please report this error to support@cognite.com and provide us with your request id: 4594f717-51c0-9708-86d3-5c5431f60cd1 and cluster name: westeurope-1"But when run it again it worked. The issue happens quite often(the number of instances in the request is close to 1000), could you have a check? Thanks. Regards,Xiaofeng 

Hi, is it possible to host a web-hook endpoint in a Cognite function?And is it a good idea?

How do I store a dataframe as an XLS in CDF files? I have to perform this within CDF functions.  

How to connect azure devops variable group and read its variable to manifest.yaml file for transformation deployment . in manifest.yaml, secret is visible. I want to hide it. transformation is getting deployed using azure devops CI/CD pipeline.

Update! See this post for an updated example notebook based on the Cognite Python SDK.We want to make it easy for you to interact with CDF Workflows through Python. Therefore we’ve created a Jupyter Notebook that demonstrates the service using Python and a simple example. Native support in the Cognite Python SDK will be released shortly. Try it out, and feel free to leave feedback or questions in the comments! Note: Download the file below and remove the .txt postfix to open the file as a Jupyter Notebook.

Hey there! We are thrilled to announce the Early Adopter Beta release of CDF Workflows.CDF Workflows is a managed process orchestration service within CDF. Using CDF Workflows, you can effectively coordinate the order and timely execution of interdependent processes such as Transformations, Functions, requests to CDF, and dynamic tasks effectively. Key Highlights Declarative Workflow Definition We have designed Workflow Orchestration with a declarative approach. Users need to specify the tasks to be orchestrated, input parameters, and the desired execution order. This intuitive approach we believe reduces complexity and lowers the bar for adaption. Orchestrating four different processes The Workflow Orchestration service supports the orchestration of four crucial types of processes: Transformations, Functions, CDF, and Dynamic Tasks. This versatility ensures that our workflows can leverage various functionalities, optimizing productivity. Versioning Capabilities Managing changes an