Product User Community

I am developing a outlier detection model. The model will be accessed by users through a Streamlit app. At the moment I am assessing whether deploying the model through the Streamlit page in CDF would be a good option. My issue is that I am not sure what the most efficient way to access and import the dataset into the app is. Would it be a good approach to create a data model and import the data from there, or extract the data straight from raw? Another idea I had was setting up a scheduled Cognite Function to download the data as a csv and letting the app access that. I need the data to be updated a few times a week and have the “import” time in the app to be fairly quick for the user experience in the app to be good.My dataset is around 1 million rows with three numeric and two non-numeric columns. For now, I have the data as a raw table in CDF.Thanks!

How can I use Machine Learning Models inside Cognite data fusion?

Hello, I am following the Learning path tutorial for CDF transformations.I’m in the step that I’m supposed to upload the csv data files to the RAW Database I’ve just created. After I select the file for upload the UI simply does nothing, poking around in DevTools I find out the upload request for the files I’m trying to upload is returning error 405 (Method Not Allowed).I am using the mocked files provided by the learning path tutorial: https://learn.cognite.com/path/data-engineer-basics-transform-and-contextualize/cdf-transformations-course/1196833 Am I doing something wrong, or is this a bug?

I am doing asset transformation asset Source Target key   categoryId   sourceDb source parentExternalId parentExternalId updatedDate   createdDate   externalId externalId isCriticalLine   description description tag   areaId   isActive   I don't have all the columns for mapping to target schemaThese are the rest target columns what to do if the mapping column is not there? and where should i add the rest additional columns from source? parentId metadata name dataSetId labels

Hi community,I have noticed a large discrepancy in runtime when running my handle function as a single Cognite Function call and when I run the handle function locally. And yes, this is the runtime of the actual CALL, not the creation/deployment of the Cognite Function. The handle function is doing some calculations on a time series with around 500 000 datapoints. Locally, it runs under 1 minute, while through Cognite Function call it takes around 9 minutes. Do you have any idea what could cause the large discrepancy?Thanks!

Hi everyone,Happy to share with you our latest features in Industrial Canvas:New features:While exploring data in Data Explorer you can now easily add any item (asset, event, 3D model etc) to a Canvas.   The entire canvas can be exported in a PDF, including the annotations. You can now customize which info/fields are relevant to you, when adding an asset or event to a canvas.      You can duplicate a canvas and copy-paste items from one canvas to another.   You can now change the order/layering of items in the Canvas. You can draw markups using the line tool when holding CMD. Several styling and bug issues have been fixed and tweaked.  Give it a try and share your feedback with us :) Thank you!

Hi Everyone,In case you haven’t discover them yet, I would like to share with your what brings January release as new features and capabilities in Industrial Canvas:Working with dataBetter visualization quality while working with time series and files. When exporting a canvas in a PDF we have eliminated the blank pages and the annotations are visible in the PDF as well. AI semantic document search Ask questions in natural language and get answers based on information found in a document. Verify answers by jumping to location of source in the document. Interpret information found in the document based on questions.CollaborationWhen I’m being tagged in a canvas, I’m getting notified via email.  Keep up to date with all your canvas versions, restore to an older version and lock a canvas for preventing changes from happening.ToolboxEnriching the toolbox with more tooling such as Industrial shapes and export your canvas in a PDF.Looking forward for your feedback!

Hi all. I’ve been trying your tutorial on How to Connect CDF to Power BI, to no avail.I am not using SUbscriptions since I do not have the permissions to write one right now. So, I was just trying to execute via Time Series API or even submit a post via curl.But the problem is with the Power BI API’s URL. Eventhough I try to make a POST, by using curl, it’ll return response 401. Can anyone give me a hand here?Do I need an admin to give me write permissions?

Hi team,I have the following questions,How does CDF handle unit of measurement. Is there a concept of storage unit system and display units system Eg; If a certain time series dataset have Celsius scale and other have Fahrenheit and kelvin scale. How does CDF store these in dataset and does it offer functionality for conversion.  If I want to build a chart with multiple time series have data with different units, does CDF offer unit conversion on the fly so that I can have everything displayed in a different units? I would like to have a comprehensive documentation for it.

Hi,I was wondering where I can find information about how to set up the permissions for a file extractor reading only files from a few sites in SharePoint, i.e., not giving the extractor access to all sites. I have solved it using the GraphAPI, but this is quite cumbersome, so I was wondering if we have documentation on how this can be set up. I’ll post my solution below for clarity on how I solved the issue:_______________________________________________________________________________You will need to create two applications. One that is used in the file extractor and one to give access to the file extractor.Register the File Extractor App: Go to the Azure Active Directory portal. Click on "Azure Active Directory" then click on "App registrations". Click on "New registration" to register a new app. Grant API Permissions: Once your app is registered, click on the app's name to go to its dashboard. Click on "API permissions" in the left panel. Click "Add a permission". Choose "Mic

When I use the search function in Data Explorer it shows information from other locations that I am not interested in.  How can I search and show results only for my location?  Thank you.  

Hi community! Dive into the insights from the World Economic Forum in Davos, where @John M. Lervik  explores the dynamic future of the energy industry, together with the panelists Dr. @Paula Doyle, Chief Digital Officer at Aker BP, Dr. Noriko Rzonca, Chief Digital Officer at Cosmo Energy, and Camilla Nilsson, CEO at Kyoto Group.The spotlight is on the critical role of trustworthy data and the enabling power of Gen AI, propelling us into gaining a competitive advantage. As the discussion unfolds, it becomes clear that each of us is integral to driving change - embracing curiosity, passion, and a commitment to be the catalysts for a transformative future.You can experience the insightful session below. We invite you to share your valuable thoughts within the community that plays a pivotal role in shaping the future.   Read the blog “Top of mind” from executives at The World Economic Forum 2024

Hi There, We have multiple CDF transformations running on schedule for every 15 minutes. Sometimes these jobs are taking more than 15 minutes and leading to Transformation failure with an error - ‘A job already runs for this transform ’. Is there any provision for us in CDF to not run a job/transformation while it is running already. 

Is there any use cases available and/or documentation with examples about configuring the PI extractor?I have found the documentation available in the web lacks information with more robust examples.Also some of the videos i have seen so far are very very basic that seemes a waste of time to watch cause they can be resumed in one phrase cause they briefly tells you what the extractor do but they dont tell how to get it to work which is what we need as engineers, a step by step guide with better examples, using fictionary data and examples could be a good starting point so we can just follow up the guides and set up instead of being crating a lot of tickets which would be reserved for troubleshooting, etc. Thank you,

I would like to make you all aware of a new tool that is now in beta: The CDF Toolkit. The toolkit comes as a python package with a CLI tool, cdf-tk, and a set of templates that can be deployed to your CDF project to get started. This initial version is focused on Infield deployment, but can be used for any type of CDF project to manage your CDF configurations and resources as code. Head over to the beta group to check out the latest small demo videos we added and test it out!

Hi,I’m trying to search for signals where I know some keywords in the name string. I’m not interested in any fuzzy matches but rather just like to see only results which contain a certain phrase. This seems not to be possible unless I activate fuzzy (which then gives to many results). Is there a way to activate “partial match” but without having fuzzy search activated?

Hi,Is it possible to display stepwise connection between datapoints in Charts rather than the standard linear interpolation?Thanks!

Hei Dear CDF usersAre you a user using charts for data science purposes / tasks or an active user of python code / jupyter notebooks ? We would love to talk to you as @Lars Moastuen would like to better understand your needs and current workflows. The team is currently working on a jupyter notebooks integration with charts and would like to better understand the needs there. Please feel to comment in the thread below if you are an user using charts for these purposes.

Greetings,I would like to get a view of which files uploaded to CDF through extractors have not been completely uploaded and only partially uploaded.In order to do this, I wanted to write a python script that compares the files sizes in the RAW metadata tables with the actual file sizes in CDF. The question is, is there a way to retrieve file size directly of a file in CDF through python SDK without having to download the files locally?Thanks in advance

As of today I think the Python SDK is only accessible via pip. While I can install a pip package into a conda environment, I don’t really like this solution. Are you planning on releasing a conda package for the Python SDK?

Hi, I'm Iñaki and I've been in the digital industry for more than a decade learning and helping customers in their Digital Transformation journey.For industries like Oil and Gas and Manufacturing it feels like I’m embarking on an adventure with every customer having its own unique complexities. But it's the constant mix of challenges, opportunities and excitement that keeps us all on our toes and motivates me. These industries are all about growth and staying in the game, and they know that embracing the digital era is the real deal.Over at Cognite, we're all about guiding organizations through this wild digital journey. Our expertise? Well, it's a combo of what we've learned from our customers and our deep dive into industry best practices. It’s like putting together a puzzle when it comes to integrating all kinds of data into these massive data lakes. Is it complicated? Yes, but it’s worth it.To really unlock the full potential of digital transformation, industries need to dig deep t

Hello CDF developers and users,Following feedback about the best practice examples provided in the documentation for how to use Parallel Retrieval / Partitions, Cognite has updated its documentation to make more explicit the limits in place for how many partitions that can be used in a request.  The document now makes explicit that a maximum of 10 partitions is supported.This guidance is found in the Pagination section of the Developer GuideIn conjunction with the updated guidance, limit enforcement will be put in place on the Assets, Events and Files APIs, to ensure predictability of performance with these APIs, effective from 1st February 2023.  Requests calling for more than 10 partitions will return a HTTP 422 ‘Unprocessable_Entity’ error message.  Similar limits will be rolled out across all CDF backend APIs, and details of these limits will be provided in their respective resource type descriptions.For any existing client applications that are calling more than 10 partitions, the

When I upload a file into Classic data model, how do I associate that file object to Flexible Data Model (FDM) object and retrieve the file object data using Graph API from FDM.   

I am looking for information related to 3d models. I would like to explore does reveal/cognite supports programmatic change of animation on 3d models which are uploaded to cdf. What i am looking for is, i have 3d model of pump and would like to change the animation speed depending on the speed i get from my api (basically real time speed which i get from my api). Does it possible ? If so could you guide me to example code to how to do that. Also the uploaded 3d model in cdf is not displaying like how the original model that was uploaded. After uploading into cdf its not displaying with colors and animations, could you tell me why it is like that ? Thanks 

Hello! Was looking around and cleaning some data, and came upon the realization that I can’t find out of you can access Cognite resources that are not contained in a dataset. That is, is there an ability to find/fetch/list Cognite resources that are “floating free”.I have tried setting  data_set_ids  And data_set_external_ids  To None, but this is the same as not having specified it.