How To...

work around the ‘Single Key

Date’ limitation for time
dependent master data
BUSINESS INFORMATION WAREHOUSE

ASAP “How to…” Paper

Applicable Releases: BW 2.0B, 2.1C

SAP (SAP America, Inc. and SAP AG) assumes no responsibility for errors or omissions in these materials.
These materials are provided “as is” without a warranty of any kind, either express or implied, including but not limited to, the
implied warranties of merchantability, fitness for a particular purpose, or non-infringement.
SAP shall not be liable for damages of any kind including without limitation direct, special, indirect, or consequential damages that
may result from the use of these materials.
SAP does not warrant the accuracy or completeness of the information, text, graphics, links or other items contained within these
materials. SAP has no control over the information that you may access through the use of hot links contained in these materials
and does not endorse your use of third party web pages nor provide any warranty whatsoever relating to third party web pages.
HOW TO ... WORK AROUND THE ‘SINGLE KEY DATE’ LIMITATION FOR TIME DEPENDENT MASTER DATA

July 2001

1998 SAP AMERICA, INC. AND SAP AG TABLE OF CONTENTS

HOW TO ... WORK AROUND THE ‘SINGLE KEY DATE’ LIMITATION FOR TIME DEPENDENT MASTER DATA

1 Introduction
The BW architecture allows for time dependent master data attributes, which are linked dynamically
to the relevant fact table characteristics (typically transaction data) at query execution time. This
approach of modeling the master data attributes externally to the InfoCubes offers great flexibility.
Data changes can be immediately reflected (after master data load and master data activation /
realignment), without even touching the InfoCubes. Moreover, historical views on the data remain
available on demand.
Just to give a few typical examples of time dependent master data attributes:
• The responsible person (owner) for a cost center.
• The annual salary of an employee.
• The retail price of an article for sale.

The specific point of time, for which the master data attributes should be selected, is identified via the
parameter ‘key date’ in the query properties screen:

(Alternatively, the key date could also be selected via a variable popup by the end-user).

While the approach outlined above covers most requirements, it does not account for ‘dynamic key
dates’, that means situations where you want to link the master data attribute time selection to some
drilldown criteria in your query (e.g. to the calendar day 0CALDAY). Including the attribute values
directly in the infocubes would help to mitigate this limitation, however with the drawback of limited
flexibility on future changes (realignments). In the following, this paper will explain a practical work-
around, at the example of a typical business requirement in the retail industry. Also, some limitations
of the work-around will be highlighted.

2001 SAP AMERICA, INC. AND SAP AG 1

HOW TO ... WORK AROUND THE ‘SINGLE KEY DATE’ LIMITATION FOR TIME DEPENDENT MASTER DATA

2 Business Scenario
In retail business, for a specific article in a BW query the respective retail price should be displayed.
The retail price of an article is generally determined by conditioning technique on the R/3 side. From
the BW perspective, the retail price has many properties of an (time-dependent) attribute:
• The retail price is specific for the article (or e.g. article / site combinations).
• The retail price for an article can change over time.
• The retail price should be tracked over time.
• The retail price might need to be restated for any point back in time.
However, as compared to master data attributes, there is one major difference in how the retail price
should be treated in reporting: It should be linked dynamically to the calendar day, i.e. for any given
point in time within the query, the valid retail price current at that specific point in time should be
displayed.

3 Solution Approach
In the following, a practical (and tested) solution approach is introduced. It takes advantage of virtual
key figures, which are populated at query execution time by the respective master data attributes,
based on the drilldown state of the query.
Note however, that there are some limitations to this approach, primarily in terms of performance:
• Master data a read via custom ABAPs at query execution time. This can have some
performance impact, especially in cases of large master data tables, and inefficient custom
ABAP developments.
• More important, once the virtual key figure is included in the query, the OLAP processor will
be required to read on the detail level, which is needed to determine the respective value. For
example, should the value be determined based on the material number and the calendar
day, the OLAP processor will make sure to read data on (at least) that detail level from the
database – even if the user did decide not to drilldown to that level, and even if aggregates
do exist on the level which the user wants to report against. (The reason for this is quite
simple: Since virtual key figures are calculated dynamically at run-time, they cannot be
included in aggregates).

2001 SAP AMERICA, INC. AND SAP AG 2

HOW TO ... WORK AROUND THE ‘SINGLE KEY DATE’ LIMITATION FOR TIME DEPENDENT MASTER DATA

4 Step By Step Solution

4.1 Include the Virtual Characteristic in the data model

The virtual key figure needs to be included in the data model, so that it can be populated by the exit,
and so it can be used in the query. Include it in all infocubes, where required. Also you can append
this field to existing (already populated) infocubes. It is also possible to include the field in a
multicube, however first it needs to be available in at least one of the underlying basic cubes.
In addition to the virtual key figure for the retail price ‘ZVRRETAIL’ another virtual key figure is
required (‘ZRPCOUNT’) here. This is to avoid the retail price to multiply by the number of times, a
specific record is found for the requested characteristic value combination (remember, retail price
should behave similar to an attribute here).

1. Create the InfoObject (of type

‘key figure’) that you want to
use as ‘retail price attribute’.
Make sure to select the correct
data type and unit of currency
(chose a unit of currency that
you can reference to in the
InfoCube).
In our example, we create the
infoobject ZVRETAIL with the
following properties:
Data type = CURR
Unit = 0LOC_CURRCY

2001 SAP AMERICA, INC. AND SAP AG 3

HOW TO ... WORK AROUND THE ‘SINGLE KEY DATE’ LIMITATION FOR TIME DEPENDENT MASTER DATA

2. Create the InfoObject (of type

‘key figure’ as well) that you
want to use as ‘counter’. Make
sure to select the correct data
type.
In our example, we create the
infoobject ZRPCOUNT with the
following properties:
Data Type = Number / DEC

3. Include the infoobjects

ZVRETAIL and ZRPCOUNT in
your InfoCube. Since you do not
populate these fields with any
data (remember, the
InfoObjects are virtual key
figures!), it is not required to
delete (and reload) the infocube
contents. Besides this, however,
you treat the virtual key figures
technically just like any other
key figure.
In our example, we are using
the infocube ZRT_C99.

2001 SAP AMERICA, INC. AND SAP AG 4

HOW TO ... WORK AROUND THE ‘SINGLE KEY DATE’ LIMITATION FOR TIME DEPENDENT MASTER DATA

4. Make sure to re-activate the

respective update rule(s). You
do not need to assign any
mapping InfoObjects to the
virtual key figures.

4.2 Maintain the Lookup Table for Master Data

In the next step, you need to maintain the lookup table, which should be read by the virtual key figure
exit. You can either use a regular table of time dependent master data attributes (i.e. the /BI0/Q…
or /BIC/Q… table), or you can alternatively use a regular transparent table created for just this
purpose in the data dictionary. In any case, you will need to make sure, of course, to populate the
table. For populating the table you have the typical tools available (data sources, custom ABAPs); the
usage of those tools should be straightforward and will not be covered in this ‘How To’ paper.

5. In our example, we create a

lookup table as a transparent
table in the data dictionary. You
can see, that it’s modeled just
like a time dependent master
data attribute table – we also
could have used such a table
directly.
Also, you can see that the retail
price is depending on the
following characteristic values:
• Sales Organization
• Distribution Channel
• Article
• Color
• Date
(And, we also included an object version
key).

2001 SAP AMERICA, INC. AND SAP AG 5

HOW TO ... WORK AROUND THE ‘SINGLE KEY DATE’ LIMITATION FOR TIME DEPENDENT MASTER DATA

4.3 Maintain the update logic for the Virtual Characteristic

In the enhancement project for the virtual characteristic, the logic for the population of the InfoObject
values is developed.

6. Via transaction CMOD, create a

customer enhancement project.
In our example, we called the
customer project LSVIRT. When
choosing the name (and short
description) for the project,
remember that you can define
only one project for all virtual
infoobjects system-wide (in case
you want to define more than
one exit). Assign a
(transportable) development
class, and save the project.

7. Assign the SAP enhancement

RSR00002 (virtual
characteristics and key figures)
to your project.

2001 SAP AMERICA, INC. AND SAP AG 6

HOW TO ... WORK AROUND THE ‘SINGLE KEY DATE’ LIMITATION FOR TIME DEPENDENT MASTER DATA

8. Via Menu >> Components, you

can display the relevant
components for this project. If
you are not familiar yet with
virtual InfoObjects, it is
absolutely essential to study
first the relevant information,
specifically the online
documentation, which can be
accessed via transaction SMOD
(enter project RSR00002).
Although the following steps
contain the complete sample
coding for our example, this
“How To” paper is no
replacement for the online
documentation.

9. Doubleclick on the term

EXIT_SAPMRSRU_001 in order to
arrive at the ABAP coding (enter
change request information
where required). From here you
can branch via ‘Goto >> Main
Program’ to the main program
SAPLXRSR.
Alternatively, you could call up
the program SAPLXRSR as well
directly via the ABAP/4 Editor, in
transaction SE38.
There will be exactly 3 areas,
where you need to do coding
within this enhancement
project:
1. INCLUDE LXRSRTOP
>> INCLUDE ZXRSRTOP
2. INCLUDE LXRSRUXX
>> INCLUDE LXRSRU02
>> INCLUDE ZXRSRU02
3. INCLUDE ZXRSRZZZ

2001 SAP AMERICA, INC. AND SAP AG 7

HOW TO ... WORK AROUND THE ‘SINGLE KEY DATE’ LIMITATION FOR TIME DEPENDENT MASTER DATA

10. In the first step of the coding *--------------------------------------

• INCLUDE ZXRSRTOP
portion, you will need to • *--------------------------------------
define several variable to be * for retail price
data: g_pos_zrt_c99_0salesorg type i.
used. In our case (for the data: g_pos_zrt_c99_0distr_chan
data: g_pos_zrt_c99_zgeneric
type
type
i.
i.
InfoCube ZRT_C99), we are data: g_pos_zrt_c99_zcolorcd
data: g_pos_zrt_c99_0calday
type i.
type i.
using variables for the data: g_pos_zrt_c99_zvrretail
data: g_pos_zrt_c99_zrpcount type i.
type i.

following InfoObjects:

 0SALESORG
 0DISTR_CHAN
 ZGENERIC
 ZCOLORCD
 0CALDAY
 ZVRRETAIL
 ZRPCOUNT

*----------------------------------------------------------------------*
11. In the next step, you decide * INCLUDE ZXRSRU02 *
*----------------------------------------------------------------------*
on whether an InfoObject
should be read or populated data: l_s_chanm type rrke_s_chanm.

during data execution. DATA: L_SUBRC1 TYPE SY-SUBRC. "execute only, if key figure really used

ZVRRETAIL and ZRPCOUNT

are the only InfoObjects to be case i_s_rkb1d-infocube.

populated, all other InfoObject when 'ZRT_C99'. "retail price

are to be read. READ TABLE I_TH_KYFNM_USED

 Also note the coding WITH KEY TABLE_LINE = 'ZVRRETAIL'
TRANSPORTING NO FIELDS.
portion, where table L_SUBRC1 = SY-SUBRC.

I_TH_KYFNM is read: This IF L_SUBRC1 = 0. "only if KYF used at all....

allows you to restrict the l_s_chanm-chanm = '0SALESORG'.

l_s_chanm-mode = rrke_c_mode-read.
usage of the enhancement append l_s_chanm to e_t_chanm.
(and the possible l_s_chanm-chanm = '0DISTR_CHAN'.
performance impacts l_s_chanm-mode = rrke_c_mode-read.
append l_s_chanm to e_t_chanm.
associated with this) to l_s_chanm-chanm = 'ZGENERIC'.
those cases, where the l_s_chanm-mode = rrke_c_mode-read.
append l_s_chanm to e_t_chanm.
virtual key figure
l_s_chanm-chanm = 'ZCOLORCD'.
‘ZVRRETAIL’ is actually l_s_chanm-mode = rrke_c_mode-read.
append l_s_chanm to e_t_chanm.
used in the query (Note:
l_s_chanm-chanm = '0CALDAY'.
this coding portion will be l_s_chanm-mode = rrke_c_mode-read.
executing at the point of append l_s_chanm to e_t_chanm.

query generation, not at append 'ZVRRETAIL' to e_t_kyfnm.

query execution time). append 'ZRPCOUNT' to e_t_kyfnm.

endif.

2001 SAP AMERICA, INC. AND SAP AG 8

HOW TO ... WORK AROUND THE ‘SINGLE KEY DATE’ LIMITATION FOR TIME DEPENDENT MASTER DATA

*----------------------------------------------------------------------*

12. In the third step of the

* INCLUDE ZXRSRZZZ *
*----------------------------------------------------------------------*

enhancement project coding, * retail price (style selling cube)

you finally define the update form user_zrt_c99 using i_s_rkb1d type rsr_s_rkb1d
logic: changing c_s_data type any.

tables: ZVIRTUAL_RETAIL.
 First, some variables need field-symbols <l_0salesorg>.
to be declared (using field field-symbols <l_0distr_chan>.
field-symbols <l_zgeneric>.
symbols) as well as any field-symbols <l_zcolorcd>.
field-symbols <l_0calday>.
database table to be field-symbols <l_zvrretail>.
field-symbols <l_zrpcount>.
referenced in the exit.
Again, if you are unsure on *break-point.

how to do this, please refer assign component g_pos_zrt_c99_0salesorg

of structure c_s_data to <l_0salesorg>.
to the documention (SMOD). assign component g_pos_zrt_c99_0distr_chan
of structure c_s_data to <l_0distr_chan>.
assign component g_pos_zrt_c99_zgeneric
 Then, the relevant data are of structure c_s_data to <l_zgeneric>.
assign component g_pos_zrt_c99_zcolorcd
read via SELECT of structure c_s_data to <l_zcolorcd>.
assign component g_pos_zrt_c99_0calday
statements. In case that of structure c_s_data to <l_0calday>.
you read directy from assign component g_pos_zrt_c99_zvrretail
of structure c_s_data to <l_zvrretail>.
master data tables, you assign component g_pos_zrt_c99_zrpcount
of structure c_s_data to <l_zrpcount>.
could also use available select * from ZVIRTUAL_RETAIL
function modules, e.g. where /BI0/0SALESORG eq <l_0salesorg>
and /BI0/0DISTR_CHAN eq <l_0distr_chan>
RSAU_READ_MASTERDATA and /BIC/ZGENERIC eq <l_zgeneric>
and /BIC/ZCOLORCD eq <l_zcolorcd>
(however note, that the and OBJVERS eq 'A'
and DATETO ge l_0calday
function modules could and DATEFROM le l_0calday.
change with any release / exit.
support package). endselect.

if sy-subrc eq 0.
 In the final step, the virtual <l_zvrretail> = ZVIRTUAL_RETAIL-/BIC/ZVRRETAIL.
<l_zrpcount> = 1.
key figures are calculated. endif.

endform.

13. Activate your enhancement

project via transaction CMOD
>> Activation.

2001 SAP AMERICA, INC. AND SAP AG 9

HOW TO ... WORK AROUND THE ‘SINGLE KEY DATE’ LIMITATION FOR TIME DEPENDENT MASTER DATA

4.4 Create and execute the Query

In the final step, you create and execute a query using the virtual key figures. In the query definition
and execution, you can treat the virtual key figures just like any other key figure. Note, that in order to
arrive at the individual retail price, you will need to create a calculated key figure (formula):
Retail Price = ZVRRETAIL / ZRPCOUNT.

Article Color Calendar Day Retail Price /$ Sales /EA

3000-99087 RED 07/01/01 295 382
06/30/01 295 354
Result 295 736
BLUE 07/01/01 199 298
06/30/01 295 75
Result 247 373
Result 271 1109

2001 SAP AMERICA, INC. AND SAP AG 10