Hello everyone, is there a feature on Quicksight to have a single graphic for a set of different of datasets and apply a sort of filter to choose the dataset for which you want to see the graphic?
If not, is there some way to build “manually” something like this?
I do not know of any easy way to do this. I was thinking about your question and it ocurred to me that if you have different datasets that you can use in the same visual they should have the same structure right? Therefore, if you could identify, in each dataset which dataset it is, and union them, you could use a filter to effectively select a dataset.
As union is not possible to be performed in QuickSight between datasets, the only way I can think of doing this is if the datasets are on an external database and you perform the union in a CustomSQL.
Another option could be to have one visual per dataset and then only show the visual relevant to the dataset you select with a parameter while having all the visuals overlapped. There is some work here since you cannot easly change the dataset related to the visual, so might have to create an analysis with all the datasets and create each visual individually or duplicate the visuals as many times as dataset you have, use the APIs for extracting and importing bundles to change the datasets used for each visual.
I think it could be useful to be able to change on the fly the dataset used by a visual as well as allowing union of datasets. At AWS, our roadmap is primarily driven by our customers. Your feedback helps us build a better service. I have tagged this as a feature request.
Hope this helps!
Hey @andres007 , thanks for your answer!
I had thought about the possibility of concatenating the datasets and having a variable to identify each one of them to filter them out and that would be a feasible solution. My only doubt is, since data would be added periodically, the dataset would become really huge and I think Quicksight would start having difficulties at handling it due to the size.
I will think about the second option you wrote. Thank you!
SPICE current capacity per dataset is now 1TB or 1b rows and will probably keep increasing, I do not know about the size of your datasets, but if you are not close to those limits I think this could be a feasible option. I am not sure if the performance will be directly impacted by the size of the dataset to be honest, based on how SPICE works, since it process data is in memory and in paralell. Maybe it is worth testing it since it looks like the easier way to implement.
Hey @andres007 I do not know I am asking the same question in another way but heres my problem:
I have 3 datasets connected to one analysis. For eg: Dataset: a,b,and c. Is it possible to use the filters of a-dataset for a visual who is using b-dataset in the same analysis?
What is the best possible solution for this?
Thanks in advance.
It is not really related to this topic, but to answer your question, you can create a control with a parameter, get the data from that parameter from one dataset column and use the value of the parameter to filter in the other visuals using different datasets.
@andres007 thanks for your quick reply mate.
Where can I find a video tutorial to carry this out? Visual Aid. Not the beloved AMAZON QUICKSIGHT PAGE.
That would be great. Thanks again.
Just to rephrase my question. I might have used the wrong terminology there. Apologies.
My use case is as follows: Both datasets are connected to the same analysis.
I have created a widget with using Dataset A. Now in the second sheet I want to create a widget which requires Value,reason from Dataset B but I also need Date from Dataset A to create this widget. Could this be done?
Would your answer still remain the same?
Can you please create a new topic with the question? Having 2 different questions in the same topic ends up being confusing for other users when looking for answers.
I believe my answer is still valid but will need more info for you on what fields are available in each visualization.