An updated version with concurrency option (thanks to @Apple for the idea and example, have a bit different implementation)
experimental.zip (857 Bytes)
1 Like
Hi @patrick
I've placed the experimental.js file in lib/modules, and renamed module from core to experimental but it's just displaying as an unknown action in the UI:
I am on Wappler 7 latest beta if that may be causing some issues?
It's shown as Unknown because Wappler doesn't know about the experimental module, the action will work correctly.
1 Like
Hey Patrick, what are all the rules for using 'parallel' and 'parallel + repeat (experimental)' in Wappler? I have several cases where I need to execute the function 'lib/my-function', but 'parallel' doesn't seem to execute it correctly.
What is the issue you have with the parallel action? With parallel all actions run simultaneous, and you can't access any result of another action that runs in parallel, that is a limitation that you get by running in parallel. The experimental repeat is a variant on the standard repeat with the difference that it runs some steps in parallel.
1 Like
I understand, the current execution was of the 'nodejs' functions using the wappler library. When I did this, the parallel execution does not work correctly, executing the functions in parallel. Instead of executing in parallel, it is executed asynchronously (default nodejs). When I changed the api logic to directly use API/Database calls (insert/update/read) the parallel execution is done correctly.
View this example 'succeeded parallel action'. 'Without 'exec (asyncAuditLog)''
Traces with succeeded parallel actions:
So look at this 'parallel execution failure' example. 'With 'exec (asyncAuditLog)''
Traces without succeeded parallel actions:
The steps inside the audit-logs will not run in parallel, they will run sequential but the database and api action step should run parallel. If you want to run the steps in the audit-logs also in parallel you have to put a parallel step in that server action.
Also not all actions are async and will not run in parallel, it works best with actions like for database, api and filesystem. Parsing of large datasets will block the thread and will not work in parallel.
It's strange — I added the 'exec' action in parallel, but after its execution, the other actions no longer run in parallel.