How can I fix this Beast Mode Window Function to display the proper value in the total row?

MosesWynn · 2021-08-16T20:48:40+00:00

There was an error rendering this rich post.

I have the following table that is using window functions in beast mode to calculate pipeline coverage.

I have these window functions for calculating the visible visible columns:

Pipeline/AOP

SUM(CASE WHEN LEFT(`Stage`, 1) != '7' THEN `Pipeline` END)
/
MIN(`AOP`)

OVER (PARTITION BY `Selling Lane`)

Late Stage/AOP

SUM(CASE WHEN LEFT(`Stage`, 1) IN ('3','4','5','6') THEN `Pipeline` END)
/
MIN(`AOP`)

OVER (PARTITION BY `Selling Lane`)

Each Selling Lane has a specific AOP and there are multiple records for each selling lane.

The issue is the total row is showing the sum for each of the Selling Lanes' coverages instead of the total aggregation. I'm assuming this is because I have the partition by clause included in there, which is needed to calculate it for each selling lane. Is there a way that I can get the total row to show the aggregate overall instead of the sum for each of the selling lanes?

Find more posts tagged with

Accepted answers

GrantSmith

The total row adds everything after beast modes have been applied. You’ll need to stack your data in an ETL using an append with your data set grouped and summed and call the rows TOTAL. Then you can use the beast mode to calculate your data for each lane and the total

jaeW_at_Onyx

I don't quite understand what you're trying to accomplish, but this is syntactically incorrect (for analyzer) ... it's fine for standard SQL. But this window function MIN(AOP) will not aggregate as expected.

SUM(CASE WHEN LEFT(`Stage`, 1) IN ('3','4','5','6') THEN `Pipeline` END)
/
MIN(`AOP`)

OVER (PARTITION BY `Selling Lane`)

Window functions in Analyzer must always include two aggregations b/c the windowing happens AFTER the initial GROUP BY clause is applied.

In the below example, by taking SUM(AOP) for each row of your displayed axis, and then taking the MIN of each SUM() you're finding the minimum aggregated AOP across all your rows.

If you did it the other way around SUM(MIN(AOP) you'd find the minimum AOP for each group and then sum the MIN(AOP) across all groups.

SUM(CASE WHEN LEFT(`Stage`, 1) IN ('3','4','5','6') THEN `Pipeline` END)
/
MIN(SUM(`AOP`)) OVER (PARTITION BY `Selling Lane`)

All comments

MosesWynn

@jaeW_at_Onyx Are you able to answer this?

MarkSnodgrass

Did you try this?

SUM(CASE WHEN LEFT(`Stage`, 1) != '7' THEN `Pipeline` END) OVER ()
/
MIN(`AOP`) OVER(PARTITION BY `SF Selling Lane`)