Hi there - we have a visualization that uses a Fac...
# gooddata-cloudm
Max Donsky
03/20/2025, 3:29 PMHi there - we have a visualization that uses a Fact in the metrics section (a score) and it displays a score for various criteria. Previously this visualized the scores correctly despite using the "sum" function, but now it appears that its adding up the scores and showing the sum for every column. Is there a way to add a fact to a metric without the SUM function but just displaying in each category as we had previously? Did something change?
Max Donsky
03/20/2025, 3:51 PMFor example - as you can see the score is the same in each column and the sum function is being used. Previously, this would show the scores per our Database records for each column and would not "Sum" the scores.
m
Mauricio Cabezas
03/20/2025, 4:08 PMHi Max, thank you for your question. Would you be so nice to share with us the Dashboard or visualisation to look it deeper? (feel free to use DM)
Mauricio Cabezas
03/20/2025, 8:51 PMHi Max, I have tried to reproduce the issue in my end by testing with Dummy data, but I was not able. Otherwise I think I can have what do you expect, now (see image)? With this I can confirm it is not issue about the
Summ
Max Donsky
03/20/2025, 8:53 PMHi Mauricio the scores are not adding properly even in your test visualization here.. they should only go up to a maximum of 25 per column
Max Donsky
03/20/2025, 8:54 PMNothing has changed in the LDM in the past few days to cause this issue so I'm not sure how this would work
m
Mauricio Cabezas
03/24/2025, 12:36 PMHi Max, Yes.. as I said there were 'Dummy Data' for that reason does not match the max(25) in each column. I would like to ask if you had the time to 'clean' and fix some warnings in the LDM?
m
Max Donsky
03/24/2025, 8:42 PMHi Mauricio. Nothing had changed in our LDM so I'm not quite sure how this randomly stopped working? Are we able to have a ticket opened on this?
m
Mauricio Cabezas
03/25/2025, 3:35 PMHi Max, It seems that the score is aggregating by ID rather than by each Reason value. I also noticed that when I remove the Date (Start*****Program) or any date from the visualization, the score is no longer aggregated by ID. Could you please confirm if the values of the score are as expected when the Date is removed? This will help us focus our efforts. Additionally, I strongly recommend keeping the Logical Data Model (LDM) clean and free of warnings, at least in the datasets involved (Users and AssessmentView). You can refer to our GDTest - D***P visualization for guidance. Also, could you clarify why there is a many-to-many (M:N) relationship between Users and AssessmentView? Would it be possible to change it to a one-to-many (1:N) relationship? Regarding opening a ticket, we still think it is not necessary at this time. Please let us know once you’ve resolved the LDM warnings in the involved datasets. Thank you for your cooperation.
m
Max Donsky
03/25/2025, 7:47 PMHi Mauricio.. I removed the dates and was not able to see a difference in the score. Again, nothing has changed in the visualizations or attributes in the drill-downs nor the LDM so I am not sure how this came about so randomly. I will check in with our team re: those connections (M:N) but I am still not having luck in seeing the scores normally. I did look at the visualization that you prepared and it appears that the scores visualize properly there when looking at it by ID, but when I add in other User level data attributes (i.e first name, last name etc.) the score gets messed up again.. any ideas here?
Max Donsky
03/25/2025, 7:55 PMOur teams response was: From purely a relational perspective Users and AssessmentView should be one to one. But I believe the many to many connection was added as a work around to not being able to see all fields in GD in the past.
m
Mauricio Cabezas
03/25/2025, 8:37 PMHi Max, please check once more in the in the test insight, Just now I noticed that the problem is when ID and Date are together, if you have only ID or only a Date, the scores are correct. Also I added more user level as you mentioned. Currently there are both, ID and Date, please try to take our one of them and you will see the correct score. I will do more test and I will keep you posted.
Mauricio Cabezas
04/01/2025, 1:46 PMHi Max, I have been checking, and I see in the dataset Users, you have Id and Userid which are not well mapped, could you please fix that mapping, and check them? Same for Ids in Assessmentsview dataset, thank you
Mauricio Cabezas
04/01/2025, 3:23 PMHi Max,
I have shared two images with you via direct message. In these images, you can see my attempt to reproduce your Logical Data Model (LDM) from scratch, which I discussed with our team. I understand it may not match exactly since I lack the details about the connections involved and the complete format and type of data sources.
As your team mentioned, there are only 1:N relationships in the images sent by DM, and the metric is functioning as expected. I would like to clarify whether M:N relationships are still necessary. I’m concerned that if we maintain these relationships, the issue may persist. Additionally, we believe the connection from StudyPermitRefusalReasons to AssessmentsView (which is M:N) may not be needed.
The primary issue lies with the LDM, and it’s important to address the mapping of different IDs involved.
Lastly, I noticed that if you include the ID from AssessmentsView in the table, the values for Score will differ. Could you check if these are the expected values?
Since there are issues with ID mapping, I believe that including the ID/AssessmentsView in the visualization will help ensure that the scores for specific dates are unique.
I hope this information is helpful.