How to aggregate CASE Statement values?

pstrand2

In my dataset, I have three different tests that aggregate the field Forecast differently. Here's a beast mode I made that works pretty well, except when I'm trying to add them all together:

CASE

WHEN Test = 1 THEN SUM (Forecast)

WHEN Test = 2 THEN SUM (DISTINCT Forecast)

WHEN Test = 3 THEN SUM (DISTINCT Forecast)

END

when I filter by each Test, it works great. But when I want all of them selected, it doesn't work at all; just defaults to one value.

Example:

Test 1 Forecast: 100

Test 2 Forecast: 350

Test 3 Forecast: 250

If I select all tests, the total should be 700. If I select Tests 1 + 2, it should be 450 and so on. However, it doesn't add these together at all. How can I do this?

GrantSmith

change your logic a bit so that you’re adding your different sums together since you want to add all the sums together based on your filters .

SUM(CASE WHEN test = 1 THEN forecast ELSE 0 END) + SUM(DISTINCT CASE WHEN test = 2 THEN forecast ELSE 0 END) + SUM(DISTINCT CASE WHEN test = 3 THEN forecast ELSE 0 END)

Sorry for the bad formatting. I’m on mobile and on vacation :)

pstrand2

awesome, that solves my problem perfectly! Thank you so much Grant!

jaeW_at_Onyx

don't use SUM(Distinct) that's too risky.

Imagine if the forecast value for a set of employees is all 50. SUM distinct would assume there was only one employee with a forecast value of 50.

The better solution is to reshape your data such that you don't have to rely on DISTINCT to deduplicate rows. have a set of rows for the employee forecast, have a second set of rows for the regional forecast and a third set of rows for the state forecast. or however you're rolling up your forecasts.