Hey Team I am seeing quite odd Rollup (Total) aggr...
# gooddata-cloudr
Raffaella Gozzini
11/03/2025, 6:39 PMHey Team I am seeing quite odd Rollup (Total) aggregated numbers, could I please share an example via DM to verify with you?
k
kapa.ai
11/03/2025, 6:39 PMHello! Yes, absolutely - you can share an example via DM to get help with the Rollup (Total) aggregation numbers you're seeing. Based on the knowledge sources, there have been several cases where users experienced unexpected Rollup (Total) values, and the support team was able to investigate by: 1. Reviewing the specific visualization or dashboard - The support team typically asks for a direct link to the report/insight to investigate using [impersonation](https://www.gooddata.com/docs/cloud/security-and-compliance/impersonation/) 2. Checking common causes - Issues with Rollup (Total) can be related to: - The number of columns in your table (there was a [bug reported](https://gooddataconnect.slack.com/archives/C04S1MSLEAW/p1756205346127909?thread_ts=1756205346.127909&cid=C04S1MSLEAW) where tables with more than 100 columns showed incorrect values in exports) - Data model structure and relationships between datasets - How metrics are defined and grouped [Rollup Totals](https://www.gooddata.com/docs/cloud/create-visualizations/visualization-types/pivot-tables/#table-totals) are "smart" aggregations that match the metric they summarize and recalculate using the underlying data rather than just summing displayed values. This is especially useful for averages and non-additive metrics like unique counts. Feel free to share your example via DM, and the support team should be able to help investigate what's causing the unexpected numbers in your specific case.
r
Raffaella Gozzini
11/03/2025, 6:39 PMWho should I send the example to?
m
Mauricio Cabezas
11/03/2025, 6:41 PMHello Raffaella, happy to help, please feel free to send the
example
Mauricio Cabezas
11/03/2025, 7:49 PMHello Raffaella, thank you for sharing the links.
I am already checking, and I will create a Test insight for troubleshooting :)
Mauricio Cabezas
11/03/2025, 9:05 PMHi Rafaella, I’ve been reviewing this, but before looking deeper, could you please clarify what result you’re expecting? Let’s focus first on the *“*Tr***** assigned” attribute. If you already have a specific total or number in mind, please share it — that will help me understand what you’d like to achieve and verify if the current behavior aligns with it.
Just to clarify, when the table is broken down by Content name, the same user can appear in several rows. This means the row totals might count some users more than once, while the Rollup (Total) recalculates across the full dataset and counts each user only once. That’s why the Rollup number differs — it’s a “non-additive” metric, as described in our documentation:
“Rollup Totals are especially useful for tables with average values, avoiding issues like averaging averages. They are also ideal for non-additive metrics like unique counts. For example, a single customer might appear in multiple rows of a table showing the number of unique customers per month, but they should only be counted once in the table total.”So before proceeding, I’d like to confirm what total you expect to see — that’ll help me determine if the current setup is correct or if we should adjust it.
15 Views