I was recently using the incremental refresh feature with SPICE datasets. These allow periodic refreshing of SPICE datasets based on a date field in the dataset and a window of historical data that is to be refreshed (e.g. the last hour).
When configuring an incremental refresh for a SPICE dataset you can choose from the following refresh frequencies:
- Every 15 minutes
- Every 30 minutes
- Hourly
- Daily
- Weekly
- Monthly
These are as shown in the QuickSight dialog when configuring an incremental refresh schedule:
Please note that it isn’t possible to configure incremental refresh for SPICE datasets that have child datasets.
The following diagram shows the refresh options when adding child datasets to a parent SPICE dataset.
When configuring full refreshes, these can run at the following frequencies:
- Hourly
- Daily
- Weekly
- Monthly
Here is the QuickSight dialog when configuring a full refresh schedule:
However, there is an option to consider if you need to apply incremental refreshes to a “parent” dataset. If you use Custom SQL in place of the child dataset, this results in a single dataset (i.e. no parent/child relationship). Therefore this dataset can be configured for incremental refreshes with all supported refresh frequencies.
Which approach you take will depend on your specific use-case and Custom SQL may not be an option. For example, if you were joining multiple child datasets and need to configure different refresh frequencies for each dataset.
I hope this helps when configuring refresh of SPICE datasets in your QuickSight solutions.
Many Thanks,
Andrew