With Celonis Process Mining 4.7, we are focussing on stability & security as well as improved analytical capabilities.
Please find the official release note presentation as well as the bug fixes attached. Additionally, this article features a technical overview of the newly introduced features below.
Before updating to Celonis Process Mining 4.7
Before updating to Celonis Process Mining 4.7, some additional steps are required. Please find the guide attached below. You can also refer to the official Update Guide.
Not adhering to the instructions below can lead to loss of configuration! These steps are not necessary in case of a fresh installation. Only application updates are affected.
Features & Improvements
Compatibility verification for system updates
With every major release, we introduce a variety of new features that often require database migrations and other structural changes. Therefore, a sequential upgrade path for major release versions is required to avoid issues.
From application version 4.7 onward, we enforce a version compatibility verification during the installation process. In case the currently installed version is not compatible with the version a user is trying to install, the installation process will cancel.
Required version to install CPM 4.7
In order to upgrade to Celonis Process Mining 4.7, a minimum installed version of 4.6.3 will be required. More information can be found in the Update Guide for version 4.7.
In case you are using a version below 4.6.3, please update Celonis according to the Update Guide of the currently installed version. The version compatibility verification is not included in installers previous to 4.7!
Multi-Event Log enabled Data Models
The Multi-Event Log technology provides the capability to analyze end-to-end business processes and their interdependencies within one analysis. It allows to connect directly linked processes and to bring indirectly linked processes into context.
The Multi-Event Log technology allows you to include multiple Event Logs with their respective case tables in one semantic Data Model. All Event Logs can be linked to each other as well as to master data via joins/foreign key relationships. A holistic analysis can then access all this data on top
Please find detailed information about the configuration here.
Compute Management UI
Previously, distributed Compute Nodes had to be managed from the config-custom.properties file. With the new release, we built a Compute Management UI around the functionality to improve usability and to eliminate the necessity of restarting the application whenever the configuration changes.
As the compute nodes are now managed in the frontend, the following properties within the config-custom.properties file will no longer affect the configuration:
Analysis - Constant Variables
A new concept of Constant Variables was introduced in the Analysis. They are similar to regular variables but can not be defined by the user.
For the beginning the following two constants were introduced:
|$CEL_USERNAME||The name of the user currently viewing/editing the Analysis.|
|$CEL_ACTIVE_FILTERS||The Filters/Selections that are currently applied to the respective Analysis. Returns the String representation of the filters, separated by ";".|
Note: Load Scripts applied on any level (Analysis, Sheet, Component) are not included!
Constants can also be forwarded to External Links, by using the new action of the Button Component.
Analysis - Opening external links with the Button Component
Additionally, there is a new action for the Button Component in the Analysis: "Open External Link"
New PQL Operators
We added a variety of new PQL Operators to grant even more flexibility in analyzing business data. Some examples to highlight:
|Replaces substrings in a column and returns the column entries with respective replacements|
|CLUSTER_STRINGS||Clusters strings in a column based on their similarity|
|LEAD & LAG||Returns the row that follows/precedes the current row by offset number of rows|
|PU_STRING_AGG||Returns the concatenation of strings from the given parent column|
|RANGE_APPEND||Creates values inside a range based on a given step size and appends them to a column|
|RUNNING_SUM||Returns the sum of the previous rows|
|TO_STRING & TO_INT||Convert INT input to STRING output and vice versa|
|Returns the greatest/least element that is not NULL from a set of values|
|TRANSIT_COLUMN||Computes transition edges between related cases from two different processes.|
|Process Index Functions|
Index functions create integer indices based on a column. The following Process Index functions are available:
Integrated PQL Example & Use Case Library
We extended the PQL Function Library by more Examples and Use Cases which show how through combining various operators, complex process related questions can be answered. The solutions are presented in interactive Analyses.
PQL - Calendar Functions Rework
- The Day/Time based calendars were merged into one calendar implementation (called WEEKDAY_CALENDAR)
- The new WEEKDAY_CALENDAR now supports shifts of true 24 hours (previously only 23h:59min)
- The factory calendar expects that start date <= end date. Invalid intervals will be ignored in the calculations
- The new Workday Calendar (previously sometimes called SAP Factory Calendar) can handle gaps in the years of the TFACS configuration
- ADD_WORKDAYS using the workday calendars configured during data load is deprecated
- ADD_WORKDAYS (AW)/WORKDAYS_BETWEEN (WB) are given an additional (optional) argument to control the behaviour
- In the reworked implementation for AW/WB, dates outside the TFACS configuration are mapped to null
- Performance was improved for all the calendar implementations
- Intersecting arbitrary calendars is now possible with the new INTERSECT function
- The timestamp of the last recorded login of a given user is now visible in the respective user profile
- Users can now be locked after a certain period of inactivity (configured in the "config-custom.properties" file via instance.usersLocker.expirationDays)
- Locked users no longer count towards the license
- API Key Authentication can now be universally disabled (configured in the "config-custom.properties" file via apiKey.authentication.active)
- Error handling was improved in several areas, e.g. PDF Export, Data Model load fails
- Improved performance of the Process Explorer in certain scenarios
The following features/operators are deprecated with Celonis Process Mining 4.7: