Graphing a [Case Statement + Window Function] calculated field

Hey everyone!

I've run into an issue while attempting to create a card to graph a calculated field that involves a case statement and a window function but I'm not sure if my implementation is wrong or if this is a Domo limitation. I have a database that records the Employee's ID, Date, Hour of Availability (24H format) during the day that the employee has input and the availability status (e.g. if Employee 123 is going to work on 10/22/2020 for 3 hours we would see a 1 when he's availabile, 0 otherwise as below):

courier_id | availability_date| Hour | Nr_Hrs_Availibility

123 | 10/22/2020 | 9 | 1

123 | 10/22/2020 | 12 | 1

123 | 10/22/2020 | 17 | 1

124 | 10/22/2020 | 10 | 1

124 | 10/22/2020 | 11 | 0

124 | 10/22/2020 | 13 | 1

What now I'm trying to do is to categorize the employees into the ones that input 8 hours or less, between 9 and 20 hours and more than 20 hours and visualize weekly what percentage of the employee base work within these three ranges. I've created the attached Calculated Field that works and populates my data table correctly but when I attempt to do a COUNT(DISTINCT 'courier_id') in order to avoid double counting the Employee IDs that are repeated and attempt to visualize it (be it a Table, Graph or anything) I get a message saying "An issue has occurred during processing. We are unable to complete the request at this time." Do you have any idea why this might be happening?

TableEx.png

CalculatedField.png

Quick Links

Accepted answers

jaeW_at_Onyx

Magic will work in a pinch. Dataset Views will work as well. you can even construct your CASE statement within the DSV.

NOTE. You CAN implement a window function inside the DSV, but be advised it can have some ... unexpected interactions in cards,

ALSO the rules of what you can and cannot do in a card extend to the rules you should follow when designing DSVs for consumption in analyzer cards.

ex. you can't filter on aggregate functions in analyzer**, so don't built an aggregation into the DSV and then expect to filter on the aggregate column in anlayzer.

same for window functions. if you implement a window function in a DSV everything used in the PARTITION or ORDER BY clause must still be an axis on the visualization.

** there is a beta that allows filtering on aggregates, but it's ... early beta and hasn't been actively developed past that point AFAIK.

All comments

AbelMM

Anyone? ?

MarkSnodgrass

@GrantSmith is highly skilled at Window functions and dates and can probably figure it out for you.

GrantSmith

Gee thanks @MarkSnodgrass ?

@AbelMM - The issue resides with your windowing function. Domo is close but isn't exactly the same as SQL windowing functions.

Because you're adding an aggregate function (COUNT distinct) it's expecting your other columns in a group by statement under the hood. To make a long story short you should be able to fix your issue by changing your window function beast mode. Replace all of your instances of SUM(`nr_hours_availability`) to SUM(SUM(`nr_hours_availability`))

To get a deeper understanding of what's happening under the hood I encourage you to view @jaeW_at_Onyx 's video at https://www.youtube.com/watch?v=eifSYZIcPzg and also check out his channel in general. He's got some great videos.

Also, for future assistance it's generally a good idea to include your beast mode as text to allow others to easily copy and paste it and test it. This will get you a quicker response.

AbelMM

Thank you so much for the video and the tips, @GrantSmith ! Unfortunately I'm still struggling a little to understand what's going on under the hood apparently. I made the suggested changes in my calculation and it seems to go a little better, since it is actually showing something for a change but not quite the expected result. I attempted both partitioning by Employee ID and by Week and also partitioning by Employee ID and ordering by Week but I think the latter yields a cumulative sum and the first works better at showing the expected result. Either way when I attempt to graph with the COUNT(DISTINCT 'courier_id'), I get two series, one for the Employee IDs and one for my calculated field. I also tried directly counting the 'courier_id' given that as per the logic since the calculated field is already grouping by Employee ID and Week, the use of DISTINCT was not needed anymore but that's not yielding better results either. Here below is the code used for the calculated field and attached an image of what I'm getting in the Card Builder. Let me know if you have any other suggestions or need any more info and again thank you so much ?

CASE
WHEN SUM(SUM(`nr_hours_availability`)) OVER(PARTITION BY `courier_id`, WEEKOFYEAR(`availability_date`,22)) <=8 THEN '<=8 HoA'
WHEN SUM(SUM(`nr_hours_availability`)) OVER(PARTITION BY `courier_id`, WEEKOFYEAR(`availability_date`,22)) >8 AND SUM(SUM(`nr_hours_availability`)) OVER(PARTITION BY `courier_id`, WEEKOFYEAR(`availability_date`,22)) <=20 THEN '<=20 HoA'
WHEN SUM(SUM(`nr_hours_availability`)) OVER(PARTITION BY `courier_id`, WEEKOFYEAR(`availability_date`,22)) >20 THEN '>20 HoA'
ELSE 'NA'
END

CardBuilder2.png

GrantSmith

Hi @AbelMM

From what I can tell your configuration should work. Have you tried graphing it as a table and validating that the beast mode is returning the correct data?

I'm also wondering if it's a Domo limitation because it's a windowing function you're attempting to plot as a series but it's treating it as a separate column and not a list of series values (typically you'll see this with multiple metric fields listed in the series)

jaeW_at_Onyx

i'm reading very quickly, but i'm pretty confident it won't work as set up.

MY ASSUMPTION OF YOUR HIGH LEVEL GOAL

A list of availabilities and the number of employees who have that xyz total availability given a certain time frame.

<8, 5

<15, 23

<21, 100

>=21, 14

If that's your goal, you can't get it using window functions because you're trying to create a window function based on employee_ID, and to use a window function you must include everything that's in the ORDER BY and the PARTITION BY clause on the card axis. ... so you can't categorize users with your CASE statement, you must show each individual user.

AbelMM

Hey @jaeW_at_Onyx

Yes, the ultimate goal is to create either a graph or table that shows the weekly number of employees who fall within each of these categories, i.e. I would need to know the total number of weekly hours input by each employee in order to count him as one of the employees that input less than 8 hours on week 40, for example. So in the end I want to end up with one column per week that is divided into 3 series that represent the ranges for the hours input in the system, in which each series is counting the employee only once based on his availability. But then again I'm including the week in one of the axis, the courier on the other one, which are the two things I'm partitioning/ordering on, so that's a bit confusing.

What would be the way to go? I'm thinking the DataSet Views would be the best way since we could create an intermediate dataset with one line per employee per week (right?) but I'm waiting for the functionality to be activated on our instance. Any other suggestions? Thanks!

jaeW_at_Onyx

Magic will work in a pinch. Dataset Views will work as well. you can even construct your CASE statement within the DSV.

NOTE. You CAN implement a window function inside the DSV, but be advised it can have some ... unexpected interactions in cards,

ALSO the rules of what you can and cannot do in a card extend to the rules you should follow when designing DSVs for consumption in analyzer cards.

ex. you can't filter on aggregate functions in analyzer**, so don't built an aggregation into the DSV and then expect to filter on the aggregate column in anlayzer.

same for window functions. if you implement a window function in a DSV everything used in the PARTITION or ORDER BY clause must still be an axis on the visualization.

** there is a beta that allows filtering on aggregates, but it's ... early beta and hasn't been actively developed past that point AFAIK.

other categories

Product Ideas
Have a Domo product enhancement idea? Submit or upvote on ideas in the Ideas Exchange.
Ideas Exchange
Suggest & vote on new features you would like to see implemented in the Domo Product.
Data Connections
Ask questions about Connectors, Workbench, Cloud Amplifier and get best practices from Domo peers
Connectors
Connectors, Custom Connectors, Writeback
Workbench
Ask questions about Workbench, a secure, client-side solution for uploading your on-premise data to Domo.
Cloud Integrations
Ask questions about Cloud Integrations and Federated Data connection to your data warehouse or lake.
Data & ETL
Ask questions about Magic ETL, SQL DataFlows, DataFusion, Dataset Views and get best practices from Domo peers
Magic ETL
Ask Magic ETL questions and get answers from Domo peers
SQL DataFlows
Ask SQL DataFlow questions and get answers from Domo peers
Datasets
Ask DataFusion and Dataset Views questions and get answers from Domo peers
Visualize & Apps
Ask questions about Beast Mode, Cards, Charting, Dashboards, Stories, Variables and get best practices from Domo peers
Dashboards
Ask Cards, Dashboards, and Stories questions and get answers from Domo peers
App Studio
Ask questions about building apps in App Studio.
Pro-code Components
Ask questions about pro-code components and Domo Bricks and get answers from Domo peers.
Charting & Analyzer
Ask questions about charting and Analyzer and get answers from Domo peers.
Calculations & Variables (Beast Mode)
Ask questions about using calculated fields and Variables (Beast Modes) in Analyzer.
AI & Data science
Ask questions about DomoAI and get answers from Domo peers.
Domo AI & AI Chat
Ask questions about AI Chat and AI assistants.
Managing AI
Ask questions about managing AI with AI Playground, AI projects, AI models, and more.
Jupyter Workspaces
Ask questions about Jupyter Workspaces, Notebooks, and file share.
Automate
Ask questions about App Framework, Workflows, Domo Bricks, Domo Developer, API and get best practices from Domo peers
Workflows
Ask questions about Task Center, building automations with Domo Workflows, and executing JavaScript or Python code with Code Engine.
Alerts
Ask questions about managing alerts in Domo and get answers from Domo peers.
Distribute
Ask questions about Domo Everywhere, Scheduled Reports, Mobile and get best practices from Domo peers
Domo Everywhere
Ask questions about embedded analytics with Domo Everywhere.
Reporting
Ask questions about Scheduled Reports, Report Builder, and Slideshow Publications.
Manage
Ask questions about Governance Administration, Approvals, Teams, Alerts, and Buzz and get best practices from Domo peers
Governance & Security
Ask questions about People, Groups, Roles, Sandbox, Activity log, Buzz, Teams, Approvals and PDP and get best practices from Domo peers
Navigation & Productivity
Ask questions about navigation, Projects & Tasks, Goals, and Buzz chat.
APIs
Ask APIs and Developer.domo.com questions and get answers from Domo peers
Add-ins & Plugins
Ask questions about plugins, Microsoft add-ins, and other third-party software integrations.
Domo Community Gallery
Watch how our Customers are using Domo to solve their complex problems.
Product Releases
Domo support and product teams are here to live-answer questions about the most recent product releases. Please post questions in this Forum board for all users to benefit (rather than submitting a support ticket).
Domo University
Questions or discussions related to Domo University, trainings and certifications
Community Forums
Getting Started
Welcome to Domo's Community Forums! You'll find everything you need to get started in this category.
Community Announcements
Get the latest from Domo's Community Team.
Social Groups
Archive
Old or outdated content that could still be found helpful.

Find more posts tagged with

Case Statement