Skip to main content
Solved

Data Slices - on new metrics that don't exist in previous snapshots

  • June 24, 2026
  • 2 replies
  • 104 views

Forum|alt.badge.img+7

We are trying to use Data Slices to do variance comparisons between versions, where some versions live in the live environment, but some versions live in snapshots. We have created new metrics (formula based) in the live environment that we want to do variance on. But these metrics do not exist in the snapshots. Without functionality to be able to create / migrate these metrics into the snapshots I feel like the data slice functionality is not very useful from a snapshot perspective. It is not reasonable to assume that you’ll have every single metric ever needed created before you start taking snapshots. 

Is there anything being considered by product on how to solve this problem?

Best answer by francois

Hi!

Have you tried creating a block with the Data Slices dimension itself? Instead of relying on the Data Slices configuration, you could just create a block from scratch and put the slices dimension as part of the structure, mapping the live data to existing blocks, and older data to imports / specific calculations?

 

Given Data Slices leverage a dimension just like any other, you could still use it as a simple dimension in itself - though it will only work in tables where all metrics share the same configuration (as we can’t mix different mappings in a table today).

Your formula could look something like NewMetric[BY: Slice.Version, Slice.Year] + ArchiveMetric[BY: Slice."Old Budget with a different structure"]

 

Snapshots are immuable by design, though you should still be able to get some effective reporting when structure changes. We’re working on long term improvements that should reduce the friction to build this!

 

Hope this helps!

 

 

2 replies

francois
Employee
Forum|alt.badge.img+13
  • Employee
  • Answer
  • June 30, 2026

Hi!

Have you tried creating a block with the Data Slices dimension itself? Instead of relying on the Data Slices configuration, you could just create a block from scratch and put the slices dimension as part of the structure, mapping the live data to existing blocks, and older data to imports / specific calculations?

 

Given Data Slices leverage a dimension just like any other, you could still use it as a simple dimension in itself - though it will only work in tables where all metrics share the same configuration (as we can’t mix different mappings in a table today).

Your formula could look something like NewMetric[BY: Slice.Version, Slice.Year] + ArchiveMetric[BY: Slice."Old Budget with a different structure"]

 

Snapshots are immuable by design, though you should still be able to get some effective reporting when structure changes. We’re working on long term improvements that should reduce the friction to build this!

 

Hope this helps!

 

 


Forum|alt.badge.img+7
  • Author
  • Author
  • July 23, 2026

Hi!

Have you tried creating a block with the Data Slices dimension itself? Instead of relying on the Data Slices configuration, you could just create a block from scratch and put the slices dimension as part of the structure, mapping the live data to existing blocks, and older data to imports / specific calculations?

 

Given Data Slices leverage a dimension just like any other, you could still use it as a simple dimension in itself - though it will only work in tables where all metrics share the same configuration (as we can’t mix different mappings in a table today).

Your formula could look something like NewMetric[BY: Slice.Version, Slice.Year] + ArchiveMetric[BY: Slice."Old Budget with a different structure"]

 

Snapshots are immuable by design, though you should still be able to get some effective reporting when structure changes. We’re working on long term improvements that should reduce the friction to build this!

 

Hope this helps!

 

 

Hi Francois,

You can’t do that, because the source of the data coming into that particular data slice is from a snapshot, the data is not accessible directly within a formula in the live environment. 

I’ve spoken with a few clients who are having this issue, and it’s becoming more and more of an issue. Even when you have simple new mapping metrics created in your live workspace they only work if the data slice source is also from the live workspace. As soon as you pull something from a snapshot, the mapping metric is no longer applied because it doesn’t exist in the snapshot. 

We are going to require a way to create net new metrics in a snapshot for data slices to be valuable. Or at the very least from the mapping metric perspective, apply the live mapping metric to all data being pulled into the table regardless of whether the slice points to the live workspace or a snapshot. 

Snapshots already do allow creation/modification of access right metrics, so the functionality seems to exist? I’m not saying we should open up snapshots to allow changes to existing metrics because you’re right, that can compromise the integrity of the snapshot. But we should allow net new metrics to be migrated from live workspaces back into snapshots.