Thursday, January 15, 2009

Note 546079 - FAQ: Background jobs in Demand Planning


This note contains frequently asked questions/answers on the subject of
the mass processing in Demannd Planning

Other terms



Q1: How should I define an "Aggregational level" in background job?
Q2: What should I put in a "Selection" field in background job?
Q3: Why can I observe a "difference" in results of interactive planning and background job?
Q4: How can I increase performance of the background job?
Q5: How should I define a macro in background job?
Q6: How should I define a release of planned demand to SNP in background job?
Q7: How should I define a transfer of planned demand to R/3 in background job?

Q1: How should I define an "Aggregation level" in background job?
A1: Select the characteristics that correspond to the aggregation
level on which you want the job to be carried out. The aggregation
level can make a big difference to the job results. When you run
the job at aggregated level (eg. you want to run a forecast
for a Sales Organization -> aggregation level = Sales Organisation)
only the aggregated data of the Sales Org. will be taken for the
forecast and than the forecast will be disaggregated to the
detailed level (eg. products) according to your aggregation/
disaggregation method defined for each key figure during definition
of planning area.
Use navigational attribues in selection and aggregation level when
you what to restrict processed data without implicit using of
Do not select characteristics which are not included in a
Selection for which you run the job.
For additional information please refer to the note 374681
and to online documentation (

Q2: What should I put in a "Selection" field in background job?
A2: You can run the job for the characteristic value combinations in a
selection(s) you created earlier in the Shuffler (Interactive
Demand Planning). It is advisable to select value(s) for the
characteristic which was meant as display characteristic in the

Q3: Why can I observe a "difference" in results of Interactive Planning
and background job?
A3: There are several issues which can make you think that there are
differencies between online and batch results.
1) The selection you entered along with the aggregation level
are used as a basis for the individual, dynamically created
selections(!), which are used for calling the forecast.
In interactive planing, the selection is used directly for the
forecast. Consequently, you might be comparing the results from
different selections.
You want to do forecast for product A. The forecasting method
is defined as HISTORY + 1 = FUTURE.
Your selection consist of 3 characteristics:
1) Sales Org. 1000
2) Division, 01 02
3) Product A B A B
History (detailed level): 10 20 30 40

The online forecast for product A looks like:
(10+30) + 1 = 41(!)
In background job with the aggregation level Sales Org., Division
and Product the forecast will be done for 2 combinations:
[10 + 1 = 11] + [30 + 1 = 31] = 42(!)

2) In both cases (online and batch) the system uses
the corrected history for the calculation of the forecast.
During batch processing the system does not save the values
for corrected history if the planning book definition for this key
figure is 'not ready for input'. During interactive planning the
system allows to save the data in this line of the planning book.
For additional info please refer to notes 357789, 372939
and to online documentation (

Q4: How can I increase a performance of background job?
A4: Because you generally do not need all data of the planning area and
because in general you need different data for different planning
processes, please maintain a specific planning book and data view
for your background job. The planning book should consist only of
the key figures which are necessary, and the time buckets in data
view should be restricted to the date you need for the job.
Horizon of the planning book must be greater than the horizons in
the forecast profile.
The time period units of the planning book and forecast profile
must be identical because you are not allowed to use planning books
with mixed period units in the background processing (note 357789).
If you set Generate log indicator for your jobs you have to delete
the log periodically. If you do not do this, performance will
deteriorate noticeably! You should regulary use the transaction
/SAPAPO/MC8K (Demand Planning -> PLanning -> Demand Planning in
Background -> Delete Job Log (note 512184). You can easily check
the size of the log by checking the table /SAPAPO/LISLOG.
For some cases (e.g. loading planning areas) you can divide a job
in smaller parts (new jobs) and run it in parallel. The same
objects cannot be processed by the parallel processes (locking
problems). There must be a separation using suitable selection
restrictions in the characteristics combinations (note 428102)

Q5: How should I define a macro in background job?
A5: Before you execute your own macro in the job, you must make sure
that you have implemented all macros that provide your own macro
with data. You must remember that standard macros in Interactive
Planning (start, level and default macro) are not automatically
executed in background job. You must choose proper aggregation
level with data to be changed by the macro.
You define an activity to run a macro for background job with
transaction SAP menu -> Demand Planning -> Environment ->
Current Settings -> Define Activities for Mass Processing
When a macro run didn't change a value (the same value already
existed) you will of course not need to store any data either.
For this reason, the message "Data were stored" will not appear
in the log. In other words, the message "Data were stored" will
only appear if new data have in fact been saved.
For additional info please refer to note 412429.

Q6: How should I define a release of planned demand to SNP in
background job?
A6: You should define an activity and a release profile (/SAPAPO/MC8S)
for mass processing to release a demand from DP to SNP.
The period of the release is defined by future and history time
time buckets profile maintained in the used dataview.
When you want to release data with offset periods you have to use
BAdI /SAPAPO/SDP_RELDATA bacause offset field of the dataview
is not taken into account for release!
After every release all SNP data of the product in the location
are deleted and only the just released data exist in SNP afterwards
You define the periodicity of the release (e.g. weeks, months) with
time buckets profile used in the dataview. When your time bucket
consist of 2 weeks and 3 months and when the time bucket is used
as history bucket and future bucket your data will be released in
this way:
M3 M2 M1 W2 W1 today W1 W2 M1 M2 M3
For additional info please refer to note 403050.
[online you can use trx: /SAPAPO/MC90 or report
/SAPAPO/RTSOUTPUT_FCST, only for small number of data(!)]

Q7: How should I define a transfer of planned demand to R/3 in
background job?
A7: You should define an activity and transfer profile (/SAPAPO/MC8U)
for mass processing to release a demand from APO DP to R/3. The
Requirements Type determines the planning strategy to be used for a
material requirement in R/3 (e.g. VSE "Planning w/o final assembly"
or VSF "Planning with final assembly"). It is not possible to
transport and save PIR on R/3 side without requirements type.
The requirements type must be transported from APO or if it is not
filled on APO side, then it must be customized on R/3 side
(material master). If you transport forecast from APO to R/3, data
on both sides must
be consistent after transport -> all transported PIR must be saved
on R/3 side or stay in outbound (inbound) queue.
When you want to change data before transfer you can use
[You can use trx /sapapo/dmp2 to transfer data from SNP to R/3]

Header Data

Release Status:Released for Customer
Released on:13.02.2003 17:42:47
Master Language:English
Priority:Recommendations/additional info
Primary Component:SCM-APO-FCS-MAP Mass Processing
Secondary Components:SCM-APO-FCS Demand Planning

SCM-APO-FCS-MAC MacroBuilder

Affected Releases


Related Notes

864950 - Performance problems in DP mass processing

512184 - Background processing: Periodically delete job log

412429 - Defining jobs with macros

403050 - Consulting note: Release from DP to SNP

398726 - DP 3.0: performance planning book/data view

390055 - Mass processing: Alerts are not generated

389641 - Background jobs in Demand Planning with CBF characteristics

374681 - Selections in mass processing

372939 - APO consulting: Compare forecast in interactive & batch mode

357789 - Consulting: Horizons forecast and planning folder

356407 - Poor performance for alert macros in batch

No comments: