DavidChurchman Coach

Here's an approach: I think you should group your metrics by 'Revenue' and 'Costs' Then, to get that Total row that is a different aggregation (% gross margin) than your value, a BeastMode that tricks the pivot table into doing that for you: Then a pivot table: One of the options on subtotals is to add padding, which acts…

You should be able to get an average in the total row by switching the aggregation to average the way you did. There's something else going on. Maybe a lot of nulls in the data? Or you accidentally clicked "Hide total" option on the measure?

Okay, I think this is what you'd want for your Test 1 Forecast beastmode: SUM(AVG(`Test`) FIXED(BY Week Ending Date)) - SUM(`Test 1`) That first half sums the average of Test, fixed by week. I use the average so we get 363K once, and then fix it by week so that it uses the average for each week, summing across those weeks.

SUM(DISTINCT) is only adding up the distinct values. In this case, all the values are the same, so it just returns 363K, where you want it to return 363K*4 for the four weeks. Depending on why you are doing a SUM(DISTINCT) rather than a SUM(), you might be able to resolve this by just doing a SUM(). Or you might need a…

Coleman is right that you'll need to share more about your BeastModes to give a full answer, but what what appears to be happening is that you are subtracting the two columns. The first column is unchanging and the second isn't. The resulting grand total is the difference between the unchanging value and the sum of the…

I think something like this would work: CASE WHEN COUNT(DISTINCT Division) FIXED (BY Customer) >1 then CONCAT(`Customer`, '-', Division) else CONCAT(Customer) END

I'll miss the cringe-bot

In AppStudio, you can set filters to persist across pages in the app, and set up links to other pages within the app (or just have people browse between pages using the drop-down). Is that similar to what you're asking for?

Flag Count is already aggregated in the BeastMode, so it doesn't let you switch the aggregation, which is why that option is greyed out, but it should still be able to aggregate in the Summary. What happens when you switch from "Use Current Value" to "Use All Values"?

So, the question is: how to have sub-totals display in-line with their category and still preserve the ability to have a total row. Short answer, no you can't do that. Long answer, you could fake it using an ETL: pre-calculate the sub-totals and totals and append the rows: Then turn off the subtotals and totals in your…

If all you are pivoting is a single value metric by two categories, like in your example, you can use the "Heat map" chart type from "Popular charts" (not the Heat Map table). That pivots your data and gives a lot of heatmap options. I also like it because it's the only table type where you can use Tooltips to get a lot of…

You could go real old-school, use a monospaced font, and pad the column so they all have the same number of characters. There might be an easier way to pad in a BeastMode, but something like this would work, where 25 is the maximum string length of a column value, and you pad it out with extra spaces ending with a period…

How is this different than a default filter view?

In your sample data, if you drop Class 2, add a total row, you can label the total row "Gross Margin", and it will appear like the other rows:

The BeastMode could be written without the Country: (sum(CASE WHEN Fruit = 'Oranges' THEN Picked ELSE 0 END) + sum(CASE WHEN Fruit = 'Apples' THEN Picked ELSE 0 END) )/ sum(CASE WHEN Fruit = 'Pineapples' THEN Picked ELSE 0 END) And the way you would "pivot" by country is by using country as a variable: You could literally…

You could use the "financial style negatives (100)" option, which is a standard way to deal with this, and uses parentheses instead of negative signs:

Yes, I would love more control over the "control" menu! I end up adding so many filter cards to my apps because I know the control menu will be too hard to navigate, which isn't the best use of screen real-estate when there's a built in, collapsible menu that should do that for me. You should submit this post as an "idea"…

Love this. It would be so much more intuitive and space efficient!

Maybe I'm misunderstanding what's in your data, but I think you just need to drop 'Class 2' from your rows. Then format your total row to have the same formatting as your other rows.

When you look at your Magic ETL's history, is the error thrown on the Python scripting tile or somewhere else? What's your Python script?

I'm sorry, I didn't read your original post very carefully. Your formula in the ETL is fine so ignore my comment about wrapping the 8s and 10s. Using your sample data, I could not recreate the error with your formulas (wrapping the shift length in a sum). Re-reading, the only thing that might be wrong is your beastmodes…

You need to aggregate shift length so the pivot table knows how to combine the cells. Wrap it in either a sum() or avg(), depending on which makes more sense for you for this case I think you might need to wrap the 8s and 10s, too

Short answer, no. Manasi suggested using text-input as the control, which is probably the easiest way for this use case. For a drop-down, once you laboriously populate the variable, it can be a pain to make sure the values you have populated for the variable on your card are the same as the ones you have in your control,…

Yup, looks better on my instance now

I see the problem with your approach. Your filter is looking for rows where the CitySelector is in the Lane, but because the CitySelector is only the Shipper_City_State, there are no rows that exist where the CitySelector is the same as the Consignee. For this approach to work, you'd need to double the length of your…

I can verify that this is happening for me, too, and that this is new behavior (and that it's annoying). As a quick 'hack' to be able to see the filtered row count, you can change the Data Preview to "sampled" with a sample size larger than your total number of rows. Then it will display the number of rows "previewed" next…

It appears logically sound. Can you give some examples of city/lane combinations that worked in your filter and ones that did not work? I suspect it's some text formatting thing. Wrapping your variables in UPPER() and TRIM() might be good, to make sure it's not a case/white space issue. If that doesn't do it, I'd need to…

One workaround on this is to use a Pivot Table instead of a Table/Mega Table. You can control which columns collapse/don't collapse by checking/unchecking the "include in subtotal" option.

HTML tables don't scroll, and pivot tables don't do HTML, so I demoed an imperfect compromise on this discussion: https://community-forums.domo.com/main/discussion/comment/95482#Comment_95482

Sadly, you can't.