Replacing a dataset : Table Vs Custom SQL Query

Hi Everyone, I have a analysis which has about 10 pages and is very extensive and tons of calculated fields.

I want to replace the data set with another data set which is identical. When the dataset is replaced all the visuals stop rendering and they have to be created again from scratch.
After some testing i have found that, when in a dataset, custom SQL query is defined, we can replace it with another custom SQL and the visuals will render properly.
But when in a dataset a table is selected and we delete it to replace it with another table, the visuals get disturbed.

Any way to change dataset when table is selected without impacting my dashboard visuals?

Hello there,

The key thing required when replacing a dataset is that you have to have the exact same field names in the new dataset and these fields all need to be the same type as they were previously. If this is not the case visuals will fail to render.

This page details our recommendations when replacing a dataset.

If after reading these recommendations, you are seeing unexpected behaviour, then please raise a support ticket and they’ll be able to investigate further.

Hello, QuickSight does not support replacing a table in dataset. Deleting a table and adding a new one will change the column ids, breaking calculated fields and filters in Analysis and Dashboard.

There are 2 work arounds you can try:

  1. Create a new dataset and use replace dataset in analysis feature. For more information, see Replacing datasets.
  2. Use public api to update PhysicalTable. For more information, see UpdateDataSet.