Incorrect Aggregation in a table view - fixed when adding a more granular field

I have an issue which I think is a possible QS bug.
I’m trying to display some data in a table and I’m missing values in a group by field.
Once I add a more specific field (like id) suddenly I see value that were missing in the a aggregation (attached 2 images showing this).
There are some filters on the visual but I don’t see how this will explain the issue.

notice how USD/MXN is missing when hedge_id is not part of the group by fields of the table

Hi @shahaf ,

Welcome to the QuickSight community!!

Can you share some detail about the filters applied in this visual?
This looks like a causal effect of the filter you applied.

Incase the filter deals with a count of records scenario, using hedge id disperses the count across each record, so the records conditions justifies.

Thanks,
Prantika

1 Like

Thanks for the warm welcome :slight_smile:

These are the applied filters:

I don’t think any of these matches the cases you are describing

Can you check the field “is one2one” ? Please add it in the visual , keep the same aggregation as used in the filter and evaluate. That is the only measure in your filter I see.

it’s not really a measure (I generally don’t often take care to fix it after QS’s automatic labeling) it’s 0 or 1 depicting some internal criteria.
Adding it does not make the missing value appear.

1 Like

I understand, but based on the icon it seems it is treated as a measure.
Can you share the exact configuration of this is one2one filter? I am trying to understand the aggregation level used in the filter.

Do you see no aggregation set for the filter?

Thanks.

1 Like

Indeed there was an aggregation set to sum
I changed it and I think now it’s solved, however I don’t understand why and what just happend.
If you can please explain or refer me to the docs to understand filter aggregation and how it relates to my issue I’d really appreciate that

Since the is one2one field, although not a measure is interpreted as measure, the aggregation option determines whether the filter to be applied before aggregation of the measure or after aggregation of the measure.

When set to sum, the value for is one2one is determined by summing up the field. For the first visual it was summed at position_symbol and hedge_id level, whereas for the second one it was calculated at position symbol level only. So the condition equals 0 was not meeting for most records.
When the aggregation is set to none, the filter is applied without aggregating the is one2one field, following the granularity at which the data is available in table.

Hope this makes sense. I do not have any resource handy to explain why this behavior. Let me check if I can find something to share across.

This page talks only about how to apply the numeric filter.