Workflow with Form and Writeback

kyle_l_smith

Hello,

My use case is as follows:

Create a workflow that allows a user to go through a list of items with "Undefined" as their value under the "Designation" column. For each undefined item, the user should be prompted with a form that allows them to see the identifier for the item and select either "Capitalization" or "Operations" as a value to replace "Undefined". When the form is complete, the value should be written back to the dataset.

I am trying to follow these videos to start working through the process:

https://www.youtube.com/watch?v=ZP7RuQ3NiUs&t=17s

https://www.youtube.com/watch?v=bOAZa2Iitpo

These videos do not include any specifics on how to write the SQL query to query the dataset and I don't find any information in the documentation. Also, I was curious if I'm even on the right course for this use case. Has anyone done anything similar successfully?

Thanks,

Kyle L Smith

Zoolander

Kyle, sounds like a great use case!

As far as the SQL query goes, it follows the same syntax as beast mode. Encapsulate your columns in backticks, strings in single quotes, etc… I'm going to try and write a sample query below, but sometimes this platform does strange encoding things.

SELECT ID, First Name as FirstName from DATA where ID < 100 and Region != 'USA'

Zoolander

The forum software eliminated the backticks, but I'm guessing you follow.

kyle_l_smith

Yes, thank you. My issue has been with the table name, where you have "DATA" and I think it might be a bug because I've demonstrated the issue to support where some names work in that section and some don't. Still waiting on a resolution.

I've found under the test button that you put the dataset's name where you have "DATA", is that correct?

Assuming that I was able to get this to work, do you think workflows and a form would be a good solution for my use case?

Zoolander

Workflows and a form should work well, yes. How many of these is someone to process? And what do you do with the response?

It may be easier using a Brick, that allows people to change the value and syncing the AppDB data to a dataset that can be a dataflow input to overwrite old values. Multiple ways to skin the cat.

As far as the table name goes, I think you can use just about anything. I've used data, dataset, table, etc…

Zoolander

Considering the source of the data and what to do with the response are important considerations.

kyle_l_smith

The source of the data is a Domo datasource created through ETL. What we end up is a list of Projects by unique IDs and I need to have a human assign either "Caapitalization" or "Operations" into a field (column) named "Designation". So, my intention is as follows:

1. A Domo datasource "Projects" is updated weekly using API connector to Jira
2. "Projects" goes through ETL to prepare the data for our reports
3. A human needs to interact to determine whether a project should have "Capitalization" or "Operations" in the "Designation" field
4. Workflow makes a list of rows with "Designation" = "Unassigned" with a "Projects" as the source
5. A form is assigned to a user that has the name of the project in question and a prompt to select "Capitalization" or "Operations"
6. When the user selects the value and completes the form, the selected value is written back to replace the value of "Unassigned" in the "Designation" field.
7. Workflow repeats for every row with a value of "Unassigned"
8. Repeat this process every week

Zoolander

OK. How many of these are there roughly every week, and are these tasks to be performed by certain individuals or by anyone of a list of people? If by certain folks, are there attributes of the task that assign it to them?

As I mentioned…it could be done two ways that I can think of. What you outlined, or alternatively using a Brick. Volume and specificity of people would likely dictate which route I take.

Super high volume and no specificity of people, brick.
Lower volume or certain people assigned certain Projects, workflow.

kyle_l_smith

3-4 specific people, could easily defined by a group. The hope would be that any of them could make the designation.

Low volume per week (less than 20, probably around 10).