Hi,

Celonis Action Flows do not really provide easy possibilities to access analyses and use data out of it.

I think you must use a workaround by triggering an ML-Workbench script that connects to your EMS instance via Pycelonis.

There you can define your Space, Package, Analysis, Component and extract the data from an OLAP table and push it into another Data Pool / Data Model.

Please find some documentation here:

Trigger ML-Script -> Trigger Machine Learning Script (celonis.com)

Pycelonis Pull Data from Analysis / OLAP: Analysis -> Pull data - PyCelonis

Pycelonis Push Data to Data Pool -> Data Pool: Push data - PyCelonis

I would love to see some more out of the box features within the Celonis Action Flow module that allows exporting data from analyses with already setup filters and OLAP tables in any kind of format (e.g excel file) or to push this data somewhere in EMS.

If you do a query in the AF with the same PQL code in the same order, the result will be exactly the OLAP.

This is our current approach. Thanks anyway :)

Thanks for that Dennis. Will try that too!

Hi @Guillermo Gost 

I’m trying to achieve output same as OLAP table in Action Flows

I’m using CSV module to get this as an attachment And converting this to readable format. in this process the date column which is supposed to be like 2023-03-16

is displaying as some random number 2093849584 and as exponential value. Can you suggest a fix for this?

Thank you

Hi Preethi,

I have done this in our current AF. You can either have a variable module before pushing the data into CSV module where you can format the date in the required format or do the conversion directly in the CSV Module

syntax:

formatDate(parseDate(“your date to be converted” , x); YYYY-MM-DD)

As Vamshi said, that's the way.

Other alternative is to convert it to string, with the format you want, when querying Celonis.

https://docs.celonis.com/en/to_string.html

Did you solve this issue? I had also difficulties to export an OLAP table but It's possible with the module Get Rows, but you need to use either fields or pql saved as KPI in a Knowledge Model