Share your thoughts on visual calculations (preview)

[jeroenterheerdt]

Hit "Reply" and let us know what you think of visual calculations! To learn more about this feature, please visit the blog post or our documentation.

Here are some areas that we'd like to hear about in particular:

• errors / issues you are experiencing
• functionality that is missing / does not work as you expected that is not listed as limitations in our documentation.
• suggestions on governance capabilities
• sorting behavior of visual and visual matrix

Thanks and we look forward to hearing your feedback!

[kleigh]

I've already mentioned a drill-through column that does not need to be visible. As well as inside the visual, this also affects exports. Some people will use white text as a work-around to achieve a "hidden" column visually, but then the data appears in the exports.

Sometimes, I want to force all rows to appear - Power BI will automatically collapse rows that are the "same" which is correct for summaries but less useful when investigating the data. Having a hidden key field would ensure rows are separate while leaving the space for columns related to the investigation.

When using columns in cell elements like "Web URL" from a related table, this does not work correctly if no columns from the related table are in the table visual and the columns are only used in the elements. Having a hidden column from the related table would force a correct model and allow the use of these fields. Currently I'm duplicating the column in my main table.

One way of sorting on multiple columns is to create a compound sort key and use that as a column. While full sorting would be better, these columns would be better hidden when they are used.

I've previously taken some notes on this based on the Ideas site. Other requests I've seen include:

• Control of columns by bookmark (better than two tables swapped by bookmark; allows optional "big" fields like a notes field or compact display).
• Control by measure (e.g. to hide all blank columns, or select by user role)
• Hiding wide key columns like UUID by default (I'm not sold on this one, seems it would be confusing)
• Columns are selected by drill level
• User controls to switch columns on and off. (Though how this would interact with field parameters is a good question...)

[jeroenterheerdt]

thank you. we are working through the export story at the moment, so this is really useful input.

[lbendlin]

Hidden is a prime example of that.

For me this discussion is similar to the one we had when windowing functions were new.  With the difference that those do allow you to reach back into the data model.

Is it already mentioned anywhere that hiding fields only seems to be possible with On-Object enabled?

[jeroenterheerdt]

that's not supposed to happen. I can hide fields fine without on-object on. Can you show a screenshot?

[lbendlin]

Oh.  You can only hide the visual calculations, not the other columns?

[jeroenterheerdt]

you can't hide columns that are used as groups, but you can hide numerical columns (implicit measures, explicit measures, calculated columns and visual calculations)

[sarab1]

How do you hide calculated columns? When creating a table visual, I am only able to hide a calculated column if it is aggregated (for example choosing Count or Average instead of Don't summarize) even if the calculated column returns a numeric value. I can hide measures and visual calculations just fine, though.

[jeroenterheerdt]

I get it. No, you can't hide an unaggregated calculated column, because that is then being used as a group not a detail. if you add an max aggregation for example you can hide it just fine. the core difference is not that this is a calculated column, it's about how you use it: a group vs a detail. You can't hide grouping columns.

[jbas]

First of all, what a great feature! I have been waiting for this, for a long time. Well done!

Right now it seems that I am not able to reference any of my measures of the model in my visual calculation.

I would like to be able to add a calculation that is only used in the specific visual. 

For instance, I have a measure [Turnover] in my model. I have a table that needs to show turnover / 1000 per region. 

Right now I have to add the [Turnover] measure to the table and hide it. Then create a visual calculation: Turnover / 1000.

Would be a lot easier, if I could just use the [Turnover] measure in my visual calculation straight away.

[jeroenterheerdt]

Hello and thanks for checking in. What you mention is in fact by design and kind of the point of this feature. You can't refer back to the model because it would bring back the complexities we're trying to "leave outside". Just bringing in the measure easily resolves this as you have discovered.

[jbas]

Thanks for your reply. I totally understand that you are trying to hide the complexity. I actually expected that you would give me the "by design" answer. 😊

However, I do feel that it has added value to be able to reference the measure from the model. Perhaps a new function could still hide the complexity, but allow expert users to use measures from the model. This way you don't have to pollute the visual with columns you have to hide.

For instance take the following visual calculation : MEASUREFROMMODEL('[Turnover]') / 1000

The MEASUREFROMMODEL function allows me to get a measure from the model. This way you still hide the complexity from basic users, but an advance user can get measure from the model.

[jeroenterheerdt]

maybe, but explaining the result of what MEASUREFROMMODEL would return is hard to do, which is the beauty of visual calcs.

[TinaM]

This visual calculation is a very cool feature, but I am not seeing the icon to hide any of the fields in my visual.

[jeroenterheerdt]

hello Tina,

thanks for checking in. The icon to hide fields should be shown in the visual calculations edit mode in the "Build a visual pane":

Please let me know if you can't find it there.

[TinaM]

Never mind...I did find it.  Sorry, I didn't have it in the edit calculation mode.

[jeroenterheerdt]

great, thanks for letting me know!

[Pim_7890]

1. Geen mogelijkheden om het format van de meting aan te passen.
   Dit moet nu aangepast worden met FORMAT().
2. Geen mogelijkheid voor voorwaardelijke opmaak van cellen.
   Dus geen mogelijkheid voor bijvoorbeeld een stoplicht-opmaak of iets dergelijks.
3. Geen mogelijkheid voor berekeningen met meerdere tabellen zoals RELATED().

[jeroenterheerdt]

hello Pim, thanks for your feedback. #1 and #2 are on the list of limitations published here.
#3 is by design. You cannot refer back to the model because you are no longer in the model. This is a core design artefact of visual calculations which makes your life easier (you don't have to worry about what makes DAX so hard because you are no longer in the model). If you want you can easily bring those columns from other tables into the visual, hide them from view so you can refer to them using visual calculations. I hope that helps.

[back_to_sql]

Hi! 

would be cool to be able to specify the orderby parameter of the running calculations. I don't think it's currently possible, or?

Other way to solve this would be to always take the sorting of the table. Right now the runningsum() function will sort by the category by default, and when you sort the table by a metric, the running sum doesn't adjust to the new sorting. 

Great addition though!

[Idrissshatila]

Hello @back_to_sql ,

from the limitations on this feature is the following : 

• You can't change the sort order for visual calculations.

check the list of limitations https://learn.microsoft.com/en-us/power-bi/transform-model/desktop-visual-calculations-overview

[CarlosAzabacheS]

It would be interesting to be able to have these calculated fields and with them create the necessary measures from the matrix we generate. Having it only in the control leaves me halfway, for example, how do I calculate the GINI, which is the result of the maximum value of a column from the matrix?