Hi Everyone,
I have an issue with tooltip which has an LAC calculation.
The pivot visual has few dimensions and then few measure like sales,profit and salesNew. In this SalesNew is a calculated field with LAC like sum(sales,[Country,City]). In Pivot visual it displays correct value for SalesNew. But in tooltip it is not showing correct value wherever it has LAC calculation.
I have attached the arena analysis for reference. Tooltip with LAC
Thanks,
Kripa
Hi @Kripa ,
This is a known limitation of how sheet tooltips interact with LAC-A calculations. Sheet tooltips don’t carry over pre-computed values from the source visual. Instead, they inherit the source visual’s filters plus an additional data-point filter for the hovered cell, and then evaluate all calculations independently in that filtered context.
From how I am reading your instance, your pivot has dimensions beyond Country and City (the ones in your LAC brackets). When you hover a specific cell, the tooltip receives a filter that narrows the data to that exact combination of ALL dimensions in the pivot, not just Country and City. Your LAC sum(sales,[Country,City]) then evaluates on this already-narrowed dataset, producing a partial sum rather than the full Country+City total you see in the pivot.
In the main pivot, the LAC-A groups by Country+City across all rows (after analysis filters), and the result is the same for every cell sharing that Country+City pair. But the tooltip’s data-point filter has already excluded rows from other dimension values before the LAC can aggregate them.
Unfortunately there’s currently no mechanism in QuickSight to pass the exact pre-computed pivot cell value into a sheet tooltip as this is confirmed as a design limitation. There’s no clean LAC workaround for this either, since PRE_FILTER would bypass the tooltip filters entirely (showing a static total regardless of which cell is hovered), and PRE_AGG still evaluates after the inherited filters.
One workaround I can think of right now if you need to display a value in the tooltip that matches what the pivot shows, the only real option right now is to pre-compute the value at the dataset/SQL level at the exact grain of Country+City, and then use MAX aggregation in the tooltip visual. I think this works only if the tooltip’s inherited filters still yield exactly one row per Country+City combination. Not ideal, but I will think a little more on this and see if I can think of any other potential workarounds for this case.