#Tableau Desktop & Web Authoring토론 중인 항목 186개

A few concrete things worth trying, since scaffolding at the full 4.5M-row grain is exactly the kind of thing that hurts performance per Tableau's own guidance on efficient calculations: 

1. Pre-aggregate before you scaffold. If rolling12/future12 is really a KPI-per-month calculation, you don't need scaffolding at the full transaction grain — aggregate to month × KPI (and whatever dimensions matter) first, either in a SQL view/CTE at the database, a Tableau Prep flow, or custom SQL in your data source. That can shrink the row count you're scaffolding against by orders of magnitude, since the window logic only needs one row per month per KPI, not one row per underlying record.
2. Do the rolling window as a table calc on the pre-aggregated result rather than a scaffold-based calc on full detail. Table calcs run after the query returns, so they're fast on a small post-aggregation result set but slow when it's still millions of rows.
3. Switch to an extract (.hyper) if you're on a live connection — extracts are generally much faster for this windowed/rolling pattern than live queries against the full 4.5M-row source.
4. If you want the heavy lifting to happen in the data source rather than Tableau's engine, an LOD expression gets compiled into the SQL as a nested select and pushed down to the database — that can outperform scaffolding since the database does the work instead of Tableau.

Official reference: Tableau's "Create Efficient Calculations" guide (

) as a non-scaffolding formula-based comparison. 

Sorry but the parameter list are set independently - i.e. they can not be change or culled based on the value selected from a different parameter - at least out of the box - there is a dasboard extension that can be used in some siturations - see https://jimdehner.com/2021/10/04/sets-and-filters-the-condition-and-top-n-tabs-what-do-they-do/

  read the entire post then focus on the 7th example - 

That said most users will use a mix of a single parameter and a relative value set or a single parameter and a set - they are easier to set up  

the choice is yours 

You could scaffold in all of your hard-coded values from a MS-excel worksheet:

Then join the above with your calculation for RC2 in the Data Source tab:

Left side is your RC2 calc, right side is the Excel worksheet Collaborations.  Then in the actual worksheet of the workbook, you'll not use your RC2 calculation to display the rows, you'll use the new Collaborations (MS-excel) worksheet values instead:

Assuming you want to count those values from RC2, that is used in the Marks Card.  As to why your results are null, well that's for you to correct.  Please do not ask us to refine your calculation; our role is to supply you with the information you originally posted/requested.  Attached is a mocked up workbook as we could not use your original workbook which is connected to Snowflake.

@Viviane Dornelas

hello and welcome back.   

Probably best to start from the beginning as it's been awhile since I last looked at the data.  As I recall, there were a number of 'weekly' reports that were posted with your question, but really only the most recent one is likely useable because it adds new data to the historical data, so the latest one is the best one. 

So, if using the one dated 2026-06-01, for June 1, 2026 for example, are you joining that worksheet called 'Complete List' with any other data from your post in Tableau Prep? Such as 'Report' worksheet from  UPDATED Published Documents Report STEM_SB?  If yes, what are you joining it on?  Or could you post the packaged workbook as .tflx file format so that we better understand what you're trying to do. 

Let's start with that question...as I wasn't able to open your originally posted Prep Flow to understand what has been done already, hence this first question.  I think, based on what I'm seeing from the .hyper file is that a lot of your data is unioned in Prep and you'll probably only need to use the most recent weekly file, not all of them. 

But we need to start with the basics first in understanding what's being joined to what in Prep...it might be simpler than originally believed.

@Curtis Bryant

Hi, measure names and measure values is not a dimension of your data. It is an artificial/built in dimension not in your data, but in the worksheet metadata.  Totals work only with dimensions/calculated dimensions in your data set. 

On the other hand, I would modify your calc to handle null values (1+null = null) (1+zn(null)=1+0=1)

{ FIXED [Student ID] : SUM(
    ZN([Grad Not Returned])
    + ZN([Grad Returned])
    + ZN([Grad Transfer])
    + ZN([Graduated])
    + ZN([Not Returned])
    + ZN([Returned])
    + ZN([Transfer])
) }

If this post resolves the question, would you be so kind to "Accept this Answer"?. This will help other users find the same answer/resolution and help community keep track of answered questions. Thank you. 

Regards, 

Diego Martinez 

Tableau Visionary and Tableau Ambassador 

Hi @Charles Rogers

Thank you for confirming. 

For example, can you check the settings of the problematic area in the workbook using the Web Editing interface on Tableau Cloud? 

How about trying to change the header color in Web Editing as a test? 

Format - Workbook

Format - Worksheet

Please note that the method for changing colors may differ depending on the location, such as Field Labels or Headers. 

However, as part of isolating the issue, please try adjusting them first.