SAP TDMS HCM : Tips and Tricks
This blog contains some information regarding TDMS HCM programs that are activity Independent. Also some tips and tricks for Performance improvement.
Program: Payroll Comparison – CNV_TDMS_HCM_PAYROLL_COMP
This Program provides the ability to do payroll result comparison,whenever you perform any of the following
- TDMS HCM data transfer
- Changes of SAP-release,
- Installation of SAP HCM support packages
- Changes in the customizing
The report can be executed in a Sender or Reciever system i.e HR system .
This Report has got two options
- Re-organize: This option allows data for inactive packages to be removed from table CNVHCM_OBJID.
- Re-Define Rules: This option allows the use of the existing packages even after upgrades. This option will redefine the existing rules in accordance with the latest SP changes, so that the existing package is reusable even after applying new SP.
The program is executed in the sender systems .This report ensures that entries required for using TDHC1 featured are maintained. In addition, this report ensures that HCM PNPCE report classes are correctly defined.
This program is called in background from the TDMS HCM activity . From this activity, certain tables that are necessary for TDMS HCM processing are filled if required. One of these tables is the CNVHCM_TR_TAB. This program can be accessed from SE38 and can be used to maintain or analyze the CNVHCM_TR_TAB table.
a) For customers on SP07, note 1526547 & 1543903 can be applied,it contains corrections for performance improvements for TEVEN table,improvements for selecting and deleting large amount of data and Post processing performance improvements.
b) Reading performance in the sender system:
Tables which are set for “Inc / Exc” value in the view V_CNVHCM_TR_TAB are read from the sender system.Any superflous or not required entries in V_CNVHCM_TR_TAB has impact on the runtime.
This is particularly important for entries for the PCL1 and PCL2 tables which can be extremely voluminous. For example if you are not interested in reading the PCL2_AR then you can unset the flag “Inc / Exc” for this entry.
c) Avoiding Memory Dumps in the PA, PA & PD and PA & PD expert mode when selecting and transferring the large amount of data for tables such as TEVEN, PCL1, PCL2, PA2001, PA2002 and so on.
-For customers on SP04 please implement Note 1554206
-For customers on SP05 to SP07 implement Note 1562976
-From SP08 this correction is available. The correction invovles a alternative logic for data selection.
This logic is briefly explianed below:
The system determines the applicable logic through the settings in table
CNVHCM_SEL_SIZE. If a particular table has no record in this table, or if the value of field MAXIMUM for this table is 0, then the old logic applies. However, if a non-zero value for field MAXIMUM is found for a particular table, then for this table the alternative logic is applied.
In this case the data is selected only in portions, which reduces the overall memory consumption.
The value of field MAXIMUM describes the number of personnel numbers
processed at a time. However, in the case of cluster tables (tables
PCLx), it describes the number of logical keys processed at a time.
We recommend to maintain large tables settings in the table
CNVHCM_SEL_SIZE. The table maintenance view V_CNVHCM_SELSIZE can be used
to maintain the entries. These settings must be available in the sender
For the following tables we recommend maintaining entries in this table:
In addition we recommend maintaining entries for the following cluster
tables. Thereby, the ‘_’ separates the name of the cluster table from
the name of the RELID.
PCL2_xx, whereby xx is a placeholder for the RELID of payroll results
(for example, PCL2_RD for the German payroll result)
In addition, we recommend maintaining entries for all large customer
tables (that is, tables named Z*).
For the size per table (field MAXIMUM of table CNVHCM_SEL_SIZE) we
recommend the following settings.
For 32 bit system:
Set 100 for all these tables
For 64 bit system with small main memory
Set 200 for all these tables
For 64 bit systems with large main memory
Set 400 for all these tables
Note that increasing this number for a particular table will improve the
runtime performance. However, in this case the system will require more
memory. In case of small systems, this may lead to an out-of-memory
dump. For that reason, if a particular table causes an out-of-memory
dump in the activity “Transfer Selection Criteria”, then it is
recommended to maintain an entry for this table in CNVHCM_SEL_SIZE with
a sufficiently small value for this table.
Delete in receiver system :
In some scenario it is required to delete redundant data existing in receiver system . In this case you can do the following
Create a new package with the selection criteria similar to the first run and execute the package till “Prepare Data Transfer” activity.
This would ensure that the target data is deleted and no data is transferred to the target system.
TDMS HCM Processes :
During execution of the Postprocessing activity for TDMS4HCM the
following is done
-Central Person creation in the Receiver system :
If the flag ‘Transfer Central Person data’ is set and concurrent
employement is not active in that case after personnel numbers are
transferred into the receiver system, the standard SAP program
RPU46CX_CENTRAL_PERSON_ONLINE is called from the Post Processing
activity. This program checks whether the new personnel numbers have
corresponding Central Person data, if this is NOT the case then a CP
record will be generated for the corresponding PERNR in HRP1000 and
-Posting Document Creation in the Receiver system:
When relevant payroll data is transferred to the receiver system and
the flag for ‘Posting document’ is set,then during the postprocessing
activity the standard report RPCIPE00 is executed in simulation mode to
create posting run in the receiver system.
The BADI CNVTDMS_HCM_TRANSFER provides methods that can be used for additional functional enhancements. These can influence the behavior of
the transfer programs and should therefore be used carefully.
For Additional information on the BADI implementation,please refer
consulting note 1558762 -TDMS4HCM: BAdI for deletion of large tables in
This contains an example of code that callable from TDHC1 feature.