Skip to main content

Celonis Product Documentation

July 2024 Release Notes

Object-centric process mining (2024-07-12)
Sort the lists of objects and events

In the Objects and Events user interface, you can now sort the lists of object types and event types by creation date or the date of the last edit, and in forward or reverse alphabetical order. Combine sorting with filtering by process, namespace, and metadata to quickly find the object type or event type you want to work with. For tips on exploring your objects and events, see The object-centric data model.

The list of object types showing the Sort icon by the Object Types heading.
Process Adherence Manager (2024-07-11)
Re-creation of current Process Adherence Manager assets required

We’ve changed how assets are created in Process Adherence Manager (PAM). Previously, you created PAM assets using object-centric data models (also known as perspectives). From 2024-07-11, you create PAM assets based on knowledge models that are defined on top of object-centric data models. Changes you’ll notice are:

  • PAM assets are consistent with Studio assets, including in Process Explorer in Views.

  • Object types have been replaced by event logs so you’ll no longer see discrepancies with event logs used elsewhere in the Celonis Platform.

  • More flexible modeling and analysis as event logs in object-centric data models use the PQL CREATE_EVENTLOG function.

  • Global filters defined in knowledge models are automatically applied to PAM assets.

  • We've removed charts and tables and Event-Object relation and Event property filters.


Any PAM assets created before 2024-07-11 are available from 2024-07-11 as read-only assets only. After 2024-09-11, they’ll be irretrievably removed. To continue using them, you’ll need to re-create any PAM assets created before 2024-07-11 between 2024-07-11 and 2024-09-11.

Re-creating assets involves reconstructing the target model for each asset and re-defining any preset filters. We envisage it will take less than 15 minutes to re-create each PAM asset. Find out how to re-create your assets here.

Object-centric process mining (2024-07-08)
Duplicate custom object types and event types

You can now make copies of a custom object type or event type, to speed up the process of modeling multiple similar objects or events. Find your original object type or event type in the list, and use the Duplicate option on the context menu (the three vertical dots) to duplicate it with a new name of your choice. The duplicate object type or event type has all the attributes of the original and all the tags that were assigned to it, and you can edit it as you want. We don't duplicate any of its relationships or its transformations. For the custom modeling instructions, see Creating custom object types and custom event types.

Object-centric process mining (2024-07-07)
Calculate throughput time for object-centric data

The PQL process function CALC_THROUGHPUT can now be used with object-centric data models. You can use the function to calculate the throughput time between events that are not directly connected to the same object type. For example, here's how to find the number of days taken between a sales order being received and its last item being shipped to a customer:

CALC_THROUGHPUT ( “o_celonis_SalesOrder”, FIRST_OCCURRENCE ["e_celonis_CreateSalesOrder"] TO LAST_OCCURRENCE ["e_celonis_CreateDeliveryItem"], DAYS ) 

For the function documentation, see CALC_THROUGHPUT.CALC_THROUGHPUT

Object-centric process mining (2024-07-06)
Possible changes to analysis results that depend on event log order

We've fixed an issue where event logs for object-centric process mining might not be correctly ordered by timestamp for event types with long include paths. This could happen when the include path for an event (the path in the schema between an included event type and the lead object) spans at least two switches in join direction (1:N:1 or N:1:N). The issue might have occurred in any of your event logs, including autogenerated and custom event logs for perspectives, and event logs you set up outside the perspective builder using the PQL function CREATE EVENTLOG.

For affected event logs, this fix corrects the order of the events, but the results of any PQL operator that depends on the event log order might be different after the fix is rolled out in your region. The process operators such as VARIANT are the most likely to be affected but others can also have this dependency. If there's an affected event log in your data model, you'll see this warning when you work with the event log in the PQL editor for Views, or in the user interface for Studio Analysis:

CREATE_EVENTLOG: The order of <event_type> events in your event log with lead object <object_type>
changed. This order changed for <number> out of <number> lead object instance(s). Consequently,
the result of any PQL operator which depends on the event log order may have changed. 
Re-validate the results of any analysis that uses this event log. For more information,
refer to the release notes.

We’ll be applying the fix to groups of regions through this month. If you work with an event log after the end of this month and don’t see the warning, the event log isn’t affected, and you don’t need to take any action.

If you do see the warning, check where the affected event logs are used in your analysis, and re-validate the analysis, especially if important numbers and heavily-used metrics are impacted. Check components and assets such as Variant Explorer, Process Explorer, and Process Adherence Manager.