How can i do cross dataset questioning for better analysis in a chat agent? what other tools to use.
I want to connect my sales data to marketing traffic data . they are separate datasets.
I added them to topic and tried asking questions. chat agent treated both datasets as asingle table and the answes were not accurate.
Hi @vasnaren,
Much like when working on other assets within Quick Sight, referring to data within different datasets can be quite tricky as the system does not understand how they link up to each other.
To have a scenario where your datasets’ can refer to each other, the best option may be to create a table join so that you are dealing with one dataset that has linked the two up and can better understand how they match up.
But there are limitations when you join multiple tables.
I see there a internal feature while making analysis and dashboards. which is while creating dashboards. filter option works acroos datasets. Why cant we use the same feature or logic for the same
Hi @vasnaren , cross-dataset queries is not possible as of now, but it is on our immediate roadmap. In the meanwhile, apart from doing joins in the dataset stage as @Brett suggested, you can use them in a dashboard and the chat agent will be able to perform joins across dashboard visuals (up to 20k rows per visual).
Noted, thank you. for now I am converting each dataset into a topic and connected them to a space and again inturn connected to a chat agent. this is working really great for bigger datasets also, without any lag.
No more queries from me.
Thank you team.