Not all datasets available to select when try to join two datasets

Hi guys, I have a dataset US_stage_services which I want to join with the US_stage_flight dataset. I click on the dataset, edit it, and then I click “add data” to select the 2nd dataset that I wish to join. As you can see it the pic not all datasets are available to select for the join. Including the one I want. Does anyone know why and how can I make them available?

Hi @Fotis_flex - are you the owner of those data sets? If yes, can you please see whether the data is refreshed or not. If the refresh is still in progress, possibly it is not allowing it.

Regards - Sanjeeb

Hi @Sanjeeb2022 . Yes I am the owner of all datasets and the refresh is completed. I have no clue what is wrong here

Hi @Fotis_flex - This should work as expected. Is there any RLS implemented these data sets? If not, this needs some detail analysis and you need to raise a ticket to AWS Customer support team. To raise the ticket, please follow the link Creating support cases and case management - AWS Support

Regards - Sanjeeb

Hi @Sanjeeb2022 . Hmm, yes there is indeed RLS to both datasets. But the same file is used to create the RLS for both datasets. Why would that be a problem?

Hi @Fotis_flex - If RLS is used in any data set, you can not use that data set in another data set. This is a limitation in QuickSight. So you may probably create the data set without RLS and then do the join and implement RLS.

Hi @David_Wong @eperts - Any suggestion on this?

Regards - Sanjeeb

Hi @Sanjeeb2022 . Thanks for info. Indeed, the RLS was the problem. It’s quite annoying though. I remove RLS, join the tables but when I turn it on again, the join fails. Seems like QS does not support joining tables that have RLS!!!

1 Like

Hi @Fotis_flex - You are correct. When you are using RLS, there is some limitation.

Regards - Sanjeeb

I’ll mark this as a feature request

A bit late to this thread but what you can do is create a data set, DatasetA, with the joined tables without RLS. Then create a new dataset, DatasetB, which uses Dataset A and implement RLS there. The joins will be hidden from DatasetB so you should be good to go.