Loading Profile...
What does "superset values are not unique" mean?
I'm trying to refresh the data in the employee's data that I had previously uploaded by using the update data routine but it is failing with a message of "superset values are not unique". I'm using the exact same CSV file from before but I changed the default team for each user in a text editor. So the new csv file has the same number of columns and row headers as the previous one. However during the initial import I did tell gooddata to ignore some of the columns and I also renamed the display name of each field to make it look nicer, I don't know if after doing that future CSV files need to use the new display names or the new amount of columns. There is also the Mapping Base for Employees attribute that shows up under the upload data page that didn't use to exist but was created when I mapped other tables to this one. Any idea what I may be doing wrong?
4
people have this question
I have this question, too!
Tell me when someone answers.
The more people who ask this question, the more it gets noticed.
The more people who ask this question, the more it gets noticed.
Create a customer community for your own organization
Plans starting at $19/month
-
Inappropriate?Hi Eric,
Here is my best guess as to what is occurring in this situation. If I understand correctly, you have already uploaded this data, and successfully performed some mappings to the data set. Now, you are trying to update this data. Upon upload, the application will then try to automatically re-create the mappings that you had done previously.
I will try to address each of your points here and hopefully the answer will fall out of this:
1) Initial upload of Employee attributes
During upload, some columns were "not imported" and some names were changed
2) Performed mappings to this data set
OK
3) Update of Employee attributes
Here is where I think the problem is. When updating an existing data set, the same number of columns (in the same order) as the original file must be present. Even if columns were "not imported" in the original upload, these columns should be included in the update. From your description, it sounds like this was done.
Another key point here is that the updated data should overwrite the existing data, not append to it. This is because of the mappings that were done previously. I am assuming that records from other data sets were mapped to this table by some user id or unique field. However, if the new data uploaded was appended to the old data, then unique ids can appear multiple times. Therefore, the application does not know which one to map to.
Potential solution
First, please make sure that the newly updated file contains no duplicates (in other words, that each record, or superset value, is unique). Secondly, please be sure to completely overwrite the existing data rather than append to it (thereby creating duplicates of existing records).
In the screenshot below, note that the "Overwrite previous uploads" box is checked. If this box is unchecked, then the new data will be appended to the existing data.
I hope this answers your problem. If you still have problems, please let us know. I would be happy to schedule a call if this isn't clear, or if this doesn't answer your question.
Regards,
Ray
support@gooddata.com
(415) 200-0194
I’m happy
The company says
this answers the question
-
Inappropriate?Hi Ray,
Thanks for the quick response. You are perfectly clear, I understand what you are saying, and your assumptions are correct in that there is a unique id field in the employees table that I am using in the mapping from two other tables (the calls and opportunities tables). However, I was already using the "overwrite previous uploads" checkbox because I was not trying to add new data, I was trying to update old data. I just did it again to double-check that fact. And there are no duplicate rows in the CSV file, it is literally the same csv file I used the first time, except I changed the team id's, but I don't have the team ID mapped in any way. I just changed each of the team ID's to be unique (1/2/3/4/5/6/7) to see if that could be the issue, but it wasn't.
This is just sample data, no corporate secrets, so I'm happy to share the CSV file or share the project. -
Inappropriate?Hi Eric,
I'm sorry we weren't able to get this to work. To look closer at this, it would be great if you could share the CSV files with me. I will load them and try to replicate the problem on my end, and hunt down any reason for this failure. Feel free to send the files to support@gooddata.com and I'll try to get back to you with what I can learn as soon as possible.
Thanks,
Ray
EMPLOYEE
