Solved

How to use the output of one workbench to trigger and be used by an automation workflow?


Badge +2

Basically, I have a workbench that produces a CSV that I want to use in other workbenches and get a result. But There doesn't seem to be a trigger that works for me in this idea.

 

Here is how the automation would look:

 

imageHow can I trigger an automation from the result of the first workbench?

How can I transfer the resulting dataset into the automation after the trigger?

 

Thank you,

Kevin

icon

Best answer by chrisatsafe 6 May 2022, 17:25

View original

4 replies

Badge +2

Hi @kbraun77​ ,

How are you intending to run the first job that produces the CSV? Is this going to be a server app, are users running this from the Run Workspace page in FME Server (or even locally on FME Desktop), or do you want this to be part of the automation but be triggered by something else? The reason I ask is because Automations have to start with a trigger like the Directory Watch Trigger you have on the Automation Canvas. In my mind, you have many options but it depends how you intend to run the first job:

  • Start the automation with the appropriate trigger to kick off the job that produces the CSV and use output keys OR the Automations Writer to pass the path or data from one workspace to another; OR
  • Use a separate automation/app/run workspace to run the first job and in the Automation, watch the directory you are writing the CSV data to. Alternatively, you can also use the Topic Notified option to trigger the automation: https://www.screencast.com/t/aNy2yAbyth
  • If running the job locally on FME Desktop, write the CSV to a network drive and use the directory watcher to watch network folder that the CSV will be written to.

Hope that gives you some ideas, if you could clarify on how you want to run the first job we can provide additional suggestions.

Badge +2

Hi @kbraun77​ ,

How are you intending to run the first job that produces the CSV? Is this going to be a server app, are users running this from the Run Workspace page in FME Server (or even locally on FME Desktop), or do you want this to be part of the automation but be triggered by something else? The reason I ask is because Automations have to start with a trigger like the Directory Watch Trigger you have on the Automation Canvas. In my mind, you have many options but it depends how you intend to run the first job:

  • Start the automation with the appropriate trigger to kick off the job that produces the CSV and use output keys OR the Automations Writer to pass the path or data from one workspace to another; OR
  • Use a separate automation/app/run workspace to run the first job and in the Automation, watch the directory you are writing the CSV data to. Alternatively, you can also use the Topic Notified option to trigger the automation: https://www.screencast.com/t/aNy2yAbyth
  • If running the job locally on FME Desktop, write the CSV to a network drive and use the directory watcher to watch network folder that the CSV will be written to.

Hope that gives you some ideas, if you could clarify on how you want to run the first job we can provide additional suggestions.

Yes the initial job is a user input server app that produces the csv, that will be outside of and before this automation but I put it in for a visual of the workflow.

 

Badge +2

Yes the initial job is a user input server app that produces the csv, that will be outside of and before this automation but I put it in for a visual of the workflow.

 

Hi @kbraun77​ ,

Got it, okay so the second bullet point should work in your case. Simply setting the destination folder for the CSV in the server app workspace to the same directory that you are watching should do the trick - this is the simplest approach.

Badge +2

Yes the initial job is a user input server app that produces the csv, that will be outside of and before this automation but I put it in for a visual of the workflow.

 

Hey, thanks again for the response.

This was a working solution, however when the tool is used by multiple users the file being overwritten causes an issue with the end results. I've tried experimenting with having the CSV exported with unique file names and then the automation reads in the newest files.

Is there an easier approach?

Reply