PyCelonis Use case

Hi @jagadish.thiru,

In the PyCelonis docs, you can find some use cases that are quite handy:

Celonis Basics - PyCelonis

Cheers,

Jan-peter

Thank you, Jan-peter Van der Steege, documentation is clear and helped.

Cheers

Jagadish Thirumalachetty E

I personally have used it to document all the tables of all the data pools of my environment, including for each column the type, length, max-min, etc...

It is not clear (at least for me) if you can use it to MODIFY the Process Model (i.e. the tables used by Studio). My feeling now is that you can not.

Hi @Guillermo Gost,

Actually, you can! At this page (Data Model: Add new data - PyCelonis) you can see how a custom created table is added to the Data Model. This is very useful for some use cases:

• Add custom data, such as mapping tables
• Adjust columns, for instance pseudonymization
• Add new columns to existing tables, for instance adding scores resulting from a machine learning model.

Happy coding!

Kind regards,

Jan-peter

Hmmm... it seems you need to reload the model.... don't you?

Of course you can do a load of only that table. Still, it fact you are modifying the tables in the Data Integrator space (i.e. the "Vertica") and then updating the Process engine data model (the one accesible from Studio)...

True, but that is always the case if the data is modified?

In other words, if you do not (partly) reload the data model, the adjusted data is not visible in Studio. Also, by reloading the data model, the Process Mining Engine will run over the activity table(s) and update the process graph (which is depicted in the Process/Variant Explorer).

Indeed it is... but I'd like to find some workaround. As I have found the time to load the Data Model can be too long for ActiveFlow scenarios.

And there could be situations where you want to update just 1 record, but then you need to reload the full table (that could be lots of records, depending the scenario)

Anyway, I think I will need to accept that's just a feature of Celonis, and find workarounds (like having a reduced process model just for Action Flow scenarios)

I see your point. For now, we have to do it with the workarounds you mentioned. I have the feeling that with the focus on event data (with the purchase of Lenses.io), the table loads will become obsolete once.