0% found this document useful (0 votes)
52 views

BI Planning NW2004s: Gregor Dieckmann Ingo Raasch NW Eim Bi DW&PL

The document discusses the concept of aggregation levels in BI planning. Aggregation levels allow modeling different data levels in a multi-provider that can be changed via manual planning or planning functions. Simple and complex aggregation levels are described as well as how aggregation levels differ from multi-providers and planning levels in BW.

Uploaded by

ashokamruth
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
52 views

BI Planning NW2004s: Gregor Dieckmann Ingo Raasch NW Eim Bi DW&PL

The document discusses the concept of aggregation levels in BI planning. Aggregation levels allow modeling different data levels in a multi-provider that can be changed via manual planning or planning functions. Simple and complex aggregation levels are described as well as how aggregation levels differ from multi-providers and planning levels in BW.

Uploaded by

ashokamruth
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOC, PDF, TXT or read online on Scribd
You are on page 1/ 7

Gregor Dieckmann Ingo Raasch NW EIM BI DW&PL

BI Planning NW2004s
Aggregation Level
26 M!r" 2#$%

Version 1.0

SAP internal

BI Planning NW2##%s

Aggregation Level

Table of Contents:
1 Introdu tion.......................................................................................................................! 2 Aggregation "e#el.............................................................................................................4
2.1 Si$%le Aggregation "e#el.................................................................................................................... 4 2.2 Co$%le& Aggregation "e#el................................................................................................................. 4 2.! Aggregation "e#el #ersus 'ultiPro#ider............................................................................................. ( 2.3.1 Navigation Attribute Fills a Characteristic in the MultiProvider.................................................5 2.4 Co$$on %ro%erties of Aggregation "e#els........................................................................................ )

! Aggregation "e#el #ersus Planning "e#el in BW*BPS.................................................) 4 Su%%ort ..............................................................................................................................+


4.1 Tra e....................................................................................................................................................... + 4.2 ABAPs and ,B Tables.......................................................................................................................... + 4.! ,e#elo%ers............................................................................................................................................. +

BI Planning NW2##%s

Aggregation Level

Seitdem man begonnen hat, die einfachsten Behauptungen zu beweisen, erweisen sich viele von ihnen als falsch. Bertran& R'ssell

1 Introdu tion
(ne )'n&amental as*ect in BI Planning is to +e a+le to mo&el the levels o) a real,time In)o-'+e .here &ata can +e change& 'sing man'al *lanning or *lanning )'nctions In BI Re*orting/Anal0sis this as*ect is not so im*ortant since en& 'sers 1cons'me1 the &ata mainl0 in )orm o) re*orts2 i e it is eno'gh to have s*ecial vie.s o) the &ata containe& in In)o-'+es B't *lanning is a+o't changing &ata in a .ell mo&ele& an& consistent .a0 Es*eciall0 the inter*la0 o) *lanning )'nctions an& man'al *lanning is im*ortant2 th's one nee&s a conce*t that s'**orts to mo&el this in a reasona+le .a0 3he conce*t 'se& in BI Planning is the Aggregation Level In short this sim*l0 is a set o) characteristics an& ke0 )ig'res Let4s give an e5am*le o) an Aggregation Level an& e5*lain .hat ha**ens .hen a ke0 )ig're val'e .ill +e change& 6via man'al *lanning or a *lanning )'nction7 Ass'me that the In)o-'+e I- contains the characteristics Pro&'ct2 Pro&'ct Gro'*2 8ersion2 9ear an& the ke0 )ig're Reven'e 3he Aggregation Level AL8L contains the same In)o(+:ects as +e)ore2 e5ce*t the characteristic Pro&'ct Produ t P$ P2 P< Produ t -rou% PG$ PG$ PG2 Version 8$ 8$ 8$ .ear 2##; 2##; 2##; /e#enue $# 2# %2

Table 1 01a ttable0 InfoCube IC Rea&ing the transaction &ata on the Aggregation Level AL8L .itho't an0 restriction .e get the )ollo.ing recor&s 6Reven'e has DB aggregation =>M7? Produ t -rou% PG$ PG2 Version 8$ 8$ .ear 2##; 2##; /e#enue <# %2

Table 2 Aggregation "e#el re ords 2,B aggregated 3e4 figure #alues5 -hanging Reven'e )rom <# to %# an& saving the ne. val'e lea&s to the )ollo.ing 6&elta7 recor& in the In)o-'+e I-? Produ t @ Produ t -rou% PG$ Version 8$ .ear 2##; /e#enue $#

I e one gets a ne. recor& in the )actta+le o) In)o-'+e I- .ith the &elta o) the ke0 )ig're val'e All characteristics o) the In)o-'+e not containe& in the Aggregation Level have the initial val'e 6not assigne&7 in the &elta recor&s Aere .e ass'me that no 1&erivation1 is 'se&2 +'t this is *art o) a conce*t 'se& in BI Planning2 calle& C6ara teristi /elations6i%s

BI Planning NW2##%s

Aggregation Level

2 Aggregation "e#el
1irst of all7 t6e Aggregation "e#el is a ne8 t4%e of InfoPro#ider7 it is defined b4 a set of 6ara teristi s and 3e4 figures of t6e underl4ing InfoPro#ider. Aggregation Levels can +e mo&ele& in the Planning Mo&eler 3ransaction /SA1 &is*la0s the e5isting Aggregation Levels in the In)oProvi&er overvie.B )rom there one can navigate to the Planning Mo&eler to maintain the Aggregation Level No. .e &escri+e the s'**orte& )eat'res o) the Aggregation Level

2.1 Si$%le Aggregation "e#el


As a +asis o) a sim*le Aggregation Level onl0 a real,time In)o-'+e can +e 'se&

2.2 Co$%le& Aggregation "e#el


3he +asis o) a com*le5 Aggregation Level is a M'ltiProvi&er that contains at least one real,time In)o-'+e +'t no sim*le Aggregation Level 3he motivation )or the e5istence o) com*le5 Aggregation Levels is to s'**ort a common mo&eling conce*t as a +asis )or *lanning )'nctions2 th's e g the *lanning )'nction maintenance is the same on +asic In)oProvi&ers an& M'ltiProvi&ers? An eas0 e5am*le is a co*0 o) act'al &ata )rom an act'al c'+e to a *lan c'+e 6the Aggregation Level is +ase& on a M'ltiProvi&er that contains the *lan an& the act'al c'+e7 In a com*le5 Aggregation Level the assignment o) characteristics )rom the +asic *rovi&er is restricte& Recall that C in technical terms C the recor&s in the M'ltiProvi&er are create& )rom recor&s o) the *art*rovi&ers 'sing the 'nion o*eration =ince *lanning is a+o't changing &ata it has to +e clear .hat it means to save the &eltas o) a change& M'ltiProvi&er recor& in the corres*on&ing *art*rovi&ers T6e $ain %oint 6ere is: 9o8 an 8e ensure t6at t6ere is a uni:ue 8a4 ba 3; Let &isc'ss this *ro+lem in the )ollo.ing e5am*le? I-DA is an act'al In)o-'+e o o -haracteristics? Pro&'ct2 Pro&'ct Gro'*2 9ear Ee0 )ig'res? Reven'e2 F'antit0

I-DP is a *lan In)o-'+e o o -haracteristics? Pro&'ct Gro'*2 9ear2 8ersion Ee0 )ig'res? Reven'e

MP is M'ltiProvi&er that contains I-DA an& I-DP o o -haracteristics? Pro&'ct2 Pro&'ct Gro'*2 9ear2 8ersion Ee0 )ig'res? Reven'e2 F'antit0

(n MP .e &e)ine an Aggregation Level AL8LDMP that contains all characteristics an& ke0 )ig'res o) MP 3he )ollo.ing t.o recor&s o) I-DA an& I-DP2 res*ectivel02 Produ t P$ Produ t -rou% PG$ Produ t -rou% PG$ .ear 2##; .ear 2##; Version 8$ /e#enue $# /e#enue <# <uantit4 $

.ill lea& to t.o recor&s in the M'ltiProvi&er MP 6or AL8LDMP7? InfoPro#ider I-DA I-DP Produ t P$ P$ Produ t -rou% PG$ PG$ .ear 2##; 2##; Version @ 8$ /e#enue $# <# <uantit4 $ #

BI Planning NW2##%s

Aggregation Level

(+serve that the )iel&s in the M'ltiProvi&er that can not +e )ille& +0 the recor&s )rom I-DA or I-DP contain the initial val'e In a&&ition the characteristic 1In)oProvi&er1 is al.a0s containe& in the M'ltiProvi&er th's on a recor& level one kno.s the 1origin1 o) the recor& B't in man'al *lanning or *lanning )'nctions ne. recor&s can +e create&B .hat sho'l& .e &o .ith the )ollo.ing recor&? InfoPro#ider I-DP Produ t P1 Produ t -rou% PG$ .ear 2##; Version 8$ /e#enue %< <uantit4 1

Aere the Pro&'ct an& the F'antit0 are not initial (ne might think the sol'tion here is to 'se the *ro:ection o) this recor& to I-DP2 +'t this .ill lea& to ver0 con)'sing e))ects2 e g &isa**earing recor&s or :'m*ing ke0 )ig're val'es in man'al *lanning 3hese recor&s are not a&missi+le $ In a PlanF'er0 the ke0 )ig're F'antit0 .ill not +e in*'t ena+le& )or cells corres*on&ing to In)oProvi&er I-DP No. ass'me2 that .e have an In)o-'+e I- .ith characteristic -ostcenter2 the M'ltiProvi&er MP contains the characteristics -ostcenter an& the 6re)erencing7 characteristic =en&er -ostcenter 6*ointing to -ostcenter7 In BW release < # it .as *ossi+le to 'se the )ollo.ing assignment? -ostcenter in MP is )ille& +0 -ostcenter in I=en&er -ostcenter in MP is also )ille& )rom -ostcenter in I-

A com*le5 Aggregation Level +ase& on this M'ltiProvi&er is not a&missa+le 6)or the same reasons as a+ove7 2

2.! Aggregation "e#el #ersus 'ultiPro#ider


3he Aggregation Level an& a M'ltiProvi&er &e)ine& on a +asic In)o-'+e have similar *ro*erties an& one might ask .h0 the Aggregation Level .as invente& )or BI Planning An iteration o) M'ltiProvi&ers is not s'**orte& in BW 6M'ltiProvi&er on a M'ltiProvi&er2 etc 7 B't )or *lanning )'nction &e)inition it *rove& 'se)'l to have a common +asis2 i e mo&eling sho'l& +e the same )or +asic In)o-'+es an& M'ltiProvi&ers Witho't the conce*t o) an Aggregation Level one .o'l& have to s'**ort the iteration o) M'ltiProvi&ers 3his .o'l& a&& a lot o) com*le5it0 also to re*orting/anal0sis scenarios M'ltiProvi&er s'**ort a ma**ing o) the In)o(+:ects o) the G*art*rovi&er4 to the In)o(+:ects o) the M'ltiProvi&er Hor *lanning one onl0 nee&s a levele& vie. o) an In)oProvi&er2 the ma**ing sho'l& +e mo&ele& onl0 once

As a res'lt2 .e invente& the Aggregation Level B't still no. one can a+'se a M'ltiProvi&er to get similar e))ects as .ith the 'se o) an Aggregation Level 3he s0stem sho'l& not allo. this

2.!.1 Na#igation Attribute 1ills a C6ara teristi in t6e 'ultiPro#ider


Ass'me an In)o-'+e contains the characteristic Pro&'ct an& Pro&'ct Gro'* In a&&ition2 Pro&'ct Gro'* is a navigation attri+'te o) Pro&'ct No. incl'&e Pro&'ct Gro'* in a M'ltiProvi&er 6+'t not Pro&'ct7 an& )ill Pro&'ct Gro'* via the navigation attri+'te o) Pro&'ct 6as containe& in the In)o-'+e7 3o create an Aggregation Level on +asis o) this M'ltiProvi&er sho'l& +e im*ossi+le 3o see that these kin&s o) M'ltiProvi&er make no sense in BI Planning consi&er the e5am*le &isc'sse& in section $ 6 the characteristic Product Group of the InfoCube *la0s no role hereI7? A)ter changing the val'e )rom <# to %# the &elta recor& )or the In)o-'+e .o'l& +e? Produ t @ Produ t -rou% @ .ear 2##; Version 8$ /e#enue $#

A mathematician might sa0? I) i? I-DP ,J MP is the canonical em+e&&ing an& P? MP ,J I-DP the canonical *ro:ection then a recor& r in MP is a&missi+le onl0 i) i6P6r77 K r2 i e r is in the image o) i We think C at least )or *lanning C these kin&s o) M'ltiProvi&er make no sense 3o check all recor&s at r'n time .o'l& +e ver0 time cons'ming2 th's it is +etter to &isa+le this 1)eat're1

BI Planning NW2##%s

Aggregation Level

th's a)ter re)resh a PlanF'er0 sho.s the ol& n'm+ers <# )or PG$2 %2 )or PG2 an& $# )or the not assigne& Pro&'ct Gro'* 6 )ille& +0 the navigation attri+'te o) the not assigne& *ro&'ct I 7 Again2 this .o'l& +e a nice e5am*le o) :'m*ing ke0 )ig're val'es that no 'ser .ill 'n&erstan&

2.4 Co$$on %ro%erties of Aggregation "e#els


3he t.o Aggregation Level t0*es share the )ollo.ing common *ro*erties/restrictions? At least one ke0 )ig're an& one characteristic have to +e containe& in the Aggregation Level >se& ke0 )ig'res m'st have DB aggregation =>M2 MIN2 MAL 3he latter t.o are 1&is*la0 onl012 i e ke0 )ig're val'es can not +e change& in man'al *lanning or via *lanning )'nctions Hor ke0 )ig'res o) t0*e Date o&er 3ime onl0 the &atat0*e GDE-4 is s'**orte& Re)erencing ke0 )ig'res are not s'**orte& in Aggregation Levels 3h's non,c'm'lative val'es an& elimination o) internal +'siness vol'me are not s'**orte& Whenever a characteristic is com*o'n&e& an& 'se& in the Aggregation Level then also all com*o'n&ing *arents have to +e incl'&e& in the Aggregation Level Whenever a ke0 )ig're is 'se& in the Aggregation Level an& this ke0 )ig're has a non,)i5e& 'nit o) meas're or c'rrenc0 the corres*on&ing 'nit characteristic has to +e containe& in the Aggregation Level Whenever a ke0 )ig're .ith e5ce*tion aggregation is 'se& in the Aggregation Level the characteristic )or the e5ce*tion aggregation has to +e containe& in the Aggregation Level2 too < 3he Aggregation Level inherits a navigation attri+'te o) the +asis In)oProvi&er i) the +ase characteristic o) the navigation attri+'te is containe& in the Aggregation Level Attention? 3he navigation attri+'tes o) the Aggregation Level are not visi+le in the Aggregation Level maintenance in the Planning Mo&eler2 +'t the0 are visi+le in the F'er0 Designer A M'ltiProvi&er can not +e the +asis )or an Aggregation Level i) one characteristic o) a *art*rovi&er )ills t.o characteristics in the M'ltiProvi&er2 c) en& o) section 2 2 I) a characteristic is constant in the +asis o) an Aggregation Level this characteristics also has to +e containe& in the Aggregation Level

=ne an use t6e >C6e 3? button in t6e Planning 'odeler: T6en t6e s4ste$ tries to add t6e ne essar4 fields to $a3e t6e Aggregation "e#el onsistent.

! Aggregation "e#el #ersus Planning "e#el in BW*BPS


3he Aggregation Level an& the Planning Level are :'st t.o instances o) the same conce*t in BI Planning an& BW,BP=2 res*ectivel0 3he +asic i&ea 6the mo&eling as*ect7 is the sameB +eca'se o) the &i))erent environments some s'**orte& )eat'res are &i))erent? An Aggregation Level has no selection *art In BI Planning the selections are mo&ele& .ith Hilters In BW,BP= Planning levels can also contain selections +'t this selection .as onl0 inten&e& to s'**ort )i5e& 6'ser in&e*en&ent7 selections Planning levels can also contain BP= varia+les 3hese .ere necessar0 to +e a+le to *arameteri"e *lanning levels 6)or re'se7 In BI Planning the Aggregation Level an& the Hilters 6selections7 are com*letel0 se*arate& conce*ts 3he Hilters can also +e *arameteri"e& +0 'sing BW varia+les An Aggregation Level can contain ke0 )ig'res .ith DB aggregation =>M2 MIN an& MAL Aggregation MIN2 MAL are onl0 s'**orte& )or &is*la0 *'r*oses2 the ke0 )ig're val'es can not +e change&? In a PlanF'er0 these ke0 )ig'res are never in*'t rea&0 In BW,BP= the &esign is some.hat more &angero's? Ee0 )ig'res .ith aggregation MIN2 MAL can +e 'se& in a Planning Level 6Planning Area7 +'t at r'n time the DB aggregation =>M is 'se&I =ince a com*le5 Aggregation Level is +ase& on a M'ltiProvi&er an assignment o) characteristics o) the *art*rovi&er to the M'ltiProvi&er is *ossi+le2 c) section 2 2 In BW,BP= M'lti Planning Areas onl0 the trivial assignment is *ossi+le
3his is onl0 tr'e in sim*le cases 3hese restrictions have their origin in the .a0 the (LAP *rocessor treats e5ce*tional aggregation B0 no. the +est .a0 to get the characteristic that .ill +e nee&e& )or the e5ce*tional aggregation is to st'&0 the im*lementation o) class l@rsr2 metho& 6e 3@aggr.

<

BI Planning NW2##%s

Aggregation Level

4 Su%%ort
Aere .e collect some in)ormation that might *rove 'se)'l )or c'stomer s'**ort

4.1 Tra e
3he +acken& 6ABAP7 im*lementation can +e trace& 6as all *arts o) the Planning Mo&eler7 'sing transaction /STT. Activate 0o'r 'ser to +e trace& in R=33 6in c'stomer s0stems this is the We+D0n*ro or Portal 'ser2 in &evelo*ment/test s0stems at =AP these 'sers ma0 +e ma**e& to a )i5e& 'ser7 3hen start the Planning Mo&eler 6e g via /SP"AN 7 an& tr0 to re*ro&'ce the c'stomer *ro+lem 3hen go +ack to R=33 an& re*la0 the trace

4.2 ABAPs and ,B Tables


Most im*ortant are the )ollo.ing s0stem o+:ects? Develo*ment class? /SP"S. -lasses? C"@/SP"S@A"V"@A 6not AL8I72 these classes im*lement the 3L(G( st')) In class C"@/SP"S@A"V"@VB/S2 metho& C9BCC one can )in& the im*lementation o) restrictions mentione& in section 2 % RH- inter)ace )or Planning Mo&eler? H'nction gro'*? /SP"S@A"V"@/1C DB ta+les? /SP"S@A"V"A.

4.! ,e#elo%ers
9o' might ask the )ollo.ing *ersons )or more lo. level in)ormation 6or check the real online &oc'mentation2 c) section % 27? Aggregation Level >I in Planning Mo&eler? Aartm't EMrner2 ABAP +acken&? Ingo Raasch

You might also like