Combined review of time references for employees who have left
Initial Situation
In the course of the introduction of SAP Information Lifecycle Management in HCM, the question of the deletion logic for data from employees who have left the company arises. In our experience, there are different requirements and solutions with regard to the reference date of the data to be deleted for employees who have left the company.
SAP is also trying to meet the requirements and against this background only recently delivered note 2783754, which is intended to make it possible to process both high-end data records and delimited data records together in one set of rules for employees who have left.
This works in principle, but does not cover all the requirements that companies make.
In the following, we would like to present various possible solutions:
Challenges & possible solutions
Of the rules created within a set of rules, only one rule can be used per run. By default, the rule that would keep the record longest is chosen. However, this does not always lead to the desired result. The problem should be illustrated using an example.
The following rules are created in a set of rules:
The first rule aims to ensure that all records are deleted as soon as their end date is 5 years ago. This is to be done for employees who have left as well as for active employees.
In addition, the data of employees who have left the company should also be destroyed if the employee left the company 10 years ago.
This is a combined request for the data of employees who have left the company. Basically, you want the data to be destroyed as soon as EITHER the end of the data set is 5 years OR the exit date is 10 years ago.
Unfortunately, the rules are not processed as desired:
Since the two data sets with the end date of May 01.05.2012st, 5 are already 10 years old, it is expected that they will be deleted. The first rule has no effect on high-end data sets, as such are only taken into account when checking for exit. However, since the exit was less than XNUMX years ago, the high-end data set must not be destroyed.
Unfortunately, only the first record for the active employee is destroyed.
This is because the two rules compete with each other and only one of them is actually executed. If several rules compete with each other, it is checked which rule would keep the data records the longest. According to the first rule, the deletion date would be 01.05.2017/31.12.2023/XNUMX; according to the second rule on XNUMX/XNUMX/XNUMX. Thus, the second rule is applied and the record is not destroyed.
However, this is not the desired result, so a solution must be found.
In the meantime, SAP has delivered a note that can be used to check for the end of a data record as well as for an exit. However, the above requirements cannot be met even with the note:
The note "2783754: Use of HCM_TERMN_DATE and HCM_END_OF_RECORD" brings with it a new condition field to distinguish between high-end and non-high-end records. This is a field of the BOOLEAN type. In this, either a Y for high-end data sets or an N for non-high-end data sets is entered.
It is now possible to add this condition field to the rules listed above:
In order to import the hint, the condition field must be created manually for each object for which the condition field is required. This also includes the maintenance of the object category-specific settings (SM34), in which the fields for value determination must be entered. This means that there is a lot of additional work to be able to import this information.
After importing the note, the following situation arises:
The actual result is now in line with expectations. Since the first infotype record of the employee who has left is no longer checked for the leaving date but for the end of the data record, it is also destroyed.
The second rule, which checks for the exit date, then applies exclusively to the high-end data sets.
In this case, the result is as expected. However, there are also data constellations where this is not the case.
Assumption: In general, all data should be deleted 10 years after the end of the data record. For those who have left, it should also be deleted 10 years after leaving.
The employee who left has not been active since December 31.12.2005, XNUMX. This results in the following expectation:
Since the end of the first two infotype records was less than 1 years ago, they must not be destroyed. For the employee who has left the company, however, it should also apply that it will be deleted as soon as the employee left the company 10 years ago. This is the case. Thus, both data records would have to be deleted by him.
With the note, however, it is only possible to distinguish between high-end and non-high data sets. This does not help in this case, since the first data record would also have to be checked for the exit.
The following rules could be formed with the hint:
However, this does not delete correctly because the non-high-end record of the employee who left is not old enough.
The note is actually only suitable for setting a different time reference for non-high-end data sets than for high-end data sets. A combined time reference, i.e. checking both the end of the data record and the exit date, for example, cannot be implemented with this.
There are two ways to still meet the above requirements:
The first checks for the end of record for all employees; the second only checks all data for the exit date for employees who have left.
It is important here that the sets of rules must be switched alternately to "productive" and "non-productive". This means an increased effort in later operation, since two preliminary, write and delete runs as well as a manual status switch of the rules must be carried out for each object.
However, the requirements would be fully covered in this way:
However, since it is very time-consuming to switch the rules every time, we recommend
the second possibility:
2nd option - a customer-specific time reference
The second - and much more elegant - variant consists in the implementation of a customer-specific time reference. The advantage here is that the logic can be implemented exactly as it is needed. The effort is only incurred once at the beginning of the implementation and there is no need to switch the rules manually. It is also possible in this way to implement the requirement using a single set of rules.
For a combined time reference that checks for both end of record and exit, the logic would look like this:
As soon as either the end of the data set or the exit is 10 years ago, the data set will be deleted.
The rules would look like this:
However, this solution is only possible if the retention period for exit and end of data record is identical.
However, if the requirement is...
...delete the data of resigned employees if either the end of the data record is 5 or the resignation is 10 years ago...
... a combined time reference is not possible. In this case, the first solution with the two different sets of rules should be prioritized.
Summary / recommendation
Before you decide on a solution, you should carefully examine your requirements. Depending on the requirement, a different solution is recommended:
Requirement A:
All data should generally be deleted x years after the end of the data record.
Issue A:
High-end records stand still
Solution A:
EITHER:
High-end data records are not considered in the individual sets of rules and are later destroyed as part of the complete deletion (HRPA_PERNR).
OR:
Importing the note "2783754: Use of HCM_TERMN_DATE and HCM_END_OF_RECORD" to implement a separate check for exit for high-end data records.
Requirement B:
Data of all employees should be deleted x years after the end of the data record.
Data of employees who have left should be deleted y years after leaving.
Problem B:
There are two competing rules for the resigned employees. The deadlines for the end of the data record and for leaving are different.
Solution B:
Two sets of rules, each with one rule (see 1st option)
Requirement C:
Data of all employees should be deleted x years after the end of the data record.
Data of employees who have left the company should be deleted x years after leaving.
Problem C:
There are two competing rules for the resigned employees. However, the deadlines for the end of the data record and for leaving are identical.
Solution C:
Implementation of a customer-specific time reference that combines the two required time references (see 2nd option)
A contribution by Sanja Pierdziwol
