BI Data Modeling: MultiProviders and InfoSets

Similar documents
INSTALLING CCRQINVOICE

Maximo Reporting: Maximo-Cognos Metadata

Essentials for IBM Cognos BI (V10.2) Day(s): 5. Overview

How to use DCI Contract Alerts

Adverse Action Letters

PAGE NAMING STRATEGIES

Introduction to Oracle Business Intelligence Enterprise Edition: OBIEE Answers 11g

Reporting Requirements Specification

Constituent Page Upgrade Utility for Blackbaud CRM

Network Rail ARMS - Asbestos Risk Management System. Training Guide for use of the Import Survey Template

Querying Data with Transact SQL

The following screens show some of the extra features provided by the Extended Order Entry screen:

Overview of Data Furnisher Batch Processing

Implementing a Data Warehouse with Microsoft SQL Server

DocAve 6 Granular Backup and Restore

TRAINING GUIDE. Overview of Lucity Spatial

Using the Swiftpage Connect List Manager

ClassFlow Administrator User Guide

Using the Swiftpage Connect List Manager

ROCK-POND REPORTING 2.1

Getting Started with the Web Designer Suite

Integrating QuickBooks with TimePro

Creating a TES Encounter/Transaction Entry Batch

Area Governors Module

HOW-TO Use SAP SUIM OR RSUSR008_009_NEW to Analysing Critical Authorisations

Extensible Query Processing in Starburst

LibrePlan at CRJ A hands-on tutorial

RISKMAN REFERENCE GUIDE TO USER MANAGEMENT (Non-Network Logins)

Milestone XProtect. NVR Installer s Guide

Admin Report Kit for Exchange Server

Performance of VSA in VMware vsphere 5

HP Server Virtualization Solution Planning & Design

Data Requirements. File Types. Timeclock

IMPORTING INFOSPHERE DATA ARCHITECT MODELS INFORMATION SERVER V8.7

Release Notes. Version

STUDIO DESIGNER. Design Projects Basic Participant

SAS Viya 3.2 Administration: Mobile Devices

Copyrights and Trademarks

Dashboard Extension for Enterprise Architect

SAP Note Plan & Consol 10.0 for NetWeaver Documentation Addendum

High Security SaaS Concept Software as a Service (SaaS) for Life Science

Date: October User guide. Integration through ONVIF driver. Partner Self-test. Prepared By: Devices & Integrations Team, Milestone Systems

$ARCSIGHT_HOME/current/user/agent/map. The files are named in sequential order such as:

Customer Upgrade Checklist

HP OpenView Performance Insight Report Pack for Quality Assurance

CLIC ADMIN USER S GUIDE

AvePoint Pipeline Pro 2.0 for Microsoft Dynamics CRM

Case Metrics Guide. January 11, 2019 Version For the most recent version of this document, visit our documentation website.

BANNER BASICS. What is Banner? Banner Environment. My Banner. Pages. What is it? What form do you use? Steps to create a personal menu

Quick Guide on implementing SQL Manage for SAP Business One

Announcing Veco AuditMate from Eurolink Technology Ltd

Valorise user guide version All rights reserved 1

A Purchaser s Guide to CondoCerts

Contents: Module. Objectives. Lesson 1: Lesson 2: appropriately. As benefit of good. with almost any planning. it places on the.

User Guide. Document Version: 1.0. Solution Version:

Best Practice: Optimizing the cube build process in SAS 9.2 Mary Simmons, SAS Institute, Cary, NC Michelle Wilkie, SAS Institute, Cary, NC

Homework: Populate and Extract Data from Your Database

EUROPEAN IP NETWORK NUMBER APPLICATION FORM & SUPPORTING NOTES

Grade 4 Mathematics Item Specification C1 TJ

VMware EVO:RAIL Customer Release Notes

GPA: Plugin for Prerequisite Checks With Solution Manager 7.1

TaiRox Mail Merge. Running Mail Merge

Design Patterns. Collectional Patterns. Session objectives 11/06/2012. Introduction. Composite pattern. Iterator pattern

Please contact technical support if you have questions about the directory that your organization uses for user management.

Custodial Integrator. Release Notes. Version 3.11 (TLM)

E2Open Multi-Collab View (MCV)

Performance of usage of MindSphere depends on the bandwidth of your internet connection.

Performance and Scalability Benchmark: Siebel CRM Release 7.7 Industry Applications on IBM eserver p690 and IBM DB2 UDB on eserver p5 570

Access 2000 Queries Tips & Techniques

Oracle CPQ Cloud Release 1. New Feature Summary

Uploading Files with Multiple Loans

ISTE-608 Test Out Written Exam and Practical Exam Study Guide

NowPrint Release Notes. Last Updated: 2/1/2012

Laboratory #13: Trigger

Exercise 4: Working with tabular data Exploring infant mortality in the 1900s

Summary. Server environment: Subversion 1.4.6

MySqlWorkbench Tutorial: Creating Related Database Tables

Microsoft Excel Extensions for Enterprise Architect

Dell EqualLogic PS Series Arrays: Expanding Windows Basic Disk Partitions

The UNIVERSITY of NORTH CAROLINA at CHAPEL HILL

IBM Cognos TM1 Web Tips and Techniques

Employee Self Service (ESS) Quick Reference Guide ESS User

Proper Document Usage and Document Distribution. TIP! How to Use the Guide. Managing the News Page

SAP First Guidance SAP BW 7.5 powered by SAP HANA. SAP HANA Appliance

Delete General Ledger Account History

Backup your Data files before you begin your cleanup! Delete General Ledger Account History. Page 1

ECE 545 Project Deliverables

Structure Query Language (SQL)

eprocurement Requisition Special Request Goods

SUPPLIER CONNECTION SUPPLIER REFERENCE GUIDE FOR LEAR SUPPLIERS

OATS Registration and User Entitlement Guide

Populate and Extract Data from Your Database

Extended Vendors lets you: Maintain vendors across multiple Sage 300 companies using the Copy Vendors functionality. o

Background Check Procedures for Sponsors

Uploading Bills, Downloading Payments, and Automating the Process. Section 1: Uploading bills

Data Miner Platinum. DataMinerPlatinum allows you to build custom reports with advanced queries. Reports > DataMinerPlatinum

DocAve 6 Granular Backup and Restore

Infrastructure Series

Data Warehouse: Introduction

An Introduction to Crescendo s Maestro Application Delivery Platform

Transcription:

BI Data Mdeling: MultiPrviders and InfSets Versin 1.0 July 4, 2006

Table f Cntents: 1 D s and Dn ts fr MultiPrviders and InfSets 2 MultiPrvider Scenaris versus Partitining Characteristics 3 Data Mdels Using InfSets 3.1 Secnd-Order Navigatinal Attributes 3.2 Data Mdel Examples fr Tempral Jins 4 Perfrmance Aspects f MultiPrviders and InfSets 5 List f Dcuments Related t MultiPrviders and InfSets Mdeling MultiPrviders and InfSets with SAP BW.dc Page 2 14.06.2012

MultiPrviders and InfSets An InfPrvider is a BI Cntent bject fr which BI queries can be created r executed in the BEx. InfPrviders are the bjects r views that are relevant fr reprting. A MultiPrvider as well as an InfSet d nt physically stre data, but display lgical views. A MultiPrvider builds up a data unin f basic InfPrviders. The cmplete data f all basic InfPrviders are available fr reprting. A MultiPrvider is interpreted at runtime as independent BI queries n each basic InfPrvider where the results are merged int a single result set. An InfSet builds up a data jin f basic InfPrviders. The valid cmbinatin f recrds frm the basic InfPrviders is determined by the jin cnditin f the InfSet. 1 D s and Dn ts fr MultiPrviders and InfSets MultiPrviders When the reprting scenari is t be extended, use a MultiPrvider as central interface between query definitin and basic InfPrviders. When anther InfPrvider is added t the MultiPrvider definitin, the technical name f a query based n the MultiPrvider remains unchanged. Use a MultiPrvider t reduce the size f the basic InfPrviders. Advantages: parallel access t underlying basic InfPrviders, lad balancing, resurce utilizatin, query pruning. Make sure that yur MultiPrvider nly retrieves data frm relevant basic InfPrviders at query runtime by Using cnstants in the design f the basic InfPrviders Using different key figures in the design f the basic InfPrviders Using characteristic 0INFOPROV when designing a query n the MultiPrvider Are yu planning t use a MultiPrvider? If s, yu have t ensure that the characteristics yu want t reprt exist in all basic InfPrviders. D nt use mre than ne nn-cumulative InfCube (InfCube with at least ne nn-cumulative key figure) because this culd lead t incrrect query results. InfSets D nt use calculatins befre aggregatin n MultiPrvider because this may lead t wrng query results. D nt cmbine basic InfPrviders having inhmgeneus data mdels in a MultiPrvider. Use the reprt-reprt interface between queries defined n the basic InfPrvider instead. Avid using nly parts f cmpund characteristics in the cnstituent basic InfPrvider f a MultiPrvider. Fr mre infrmatin, see SAP nte 702542. D nt use mre than 10 InfPrviders in ne InfSet. It is better t create multiple InfSets depending n reprting needs. D nt use mre than 10 jins in ne InfSet (especially if yu expect high a data vlume). InfSet queries can be used fr DataStre bjects withut the activated BEx Reprting indicatr. See als the Perfrmance Aspects sectin f this dcument. D nt use calculatins befre aggregatin n InfSet because this may lead t wrng query results. If there are InfSets with time-dependent master data, d nt restrict the data by the fields Valid frm (0DATEFROM) and Valid t (0DATETO). See als an example in the Data Mdels using InfSets sectin f this dcument. Mdeling MultiPrviders and InfSets with SAP BW.dc Page 3 14.06.2012

2 MultiPrvider Scenaris versus Partitining Characteristics During the BI data mdeling phase, there are many key figures such as: Plan / Actual / Target Csts, Planned Sales Amunt, Actual Sales Amunt, Statistical / Actual Amunt, Ttal Sales Amunt in Lcal Currency, Ttal Sales Amunt in Dcument Currency. In the fllwing sectin, three different scenaris are described, which can be used as a BI data mdel fr such key figures. In cnclusin, we valuate the three scenaris in a cmparing matrix with respect t their pssible use cases. Scenari 1: InfCube Cntains All Key Figures (key figure data mdel) All key figures f the data mdel are defined as InfObject and are added t the fact table f ne InfCube. Examples The 3 key figures Plan Csts f Versin 1, Plan Csts f Versin 2, and Actual Csts are included in ne InfCube. The 2 key figures Ttal Sales Amunt in Lcal Currency and Ttal Sales Amunt in Dcument Currency are included in ne InfCube. Several detailed views crrespnding t the cmmn characteristics Sales Order, Custmer, and Prduct are included in ne InfCube: The key figures Sales Order Quantity and Sales Order Price refer t the detail characteristics Sales Order Date and Sales Persn. The key figures Delivered Quantity and Delivery Price refer t the detail characteristics Delivery Date and Delivery Persn. The key figures Billed Quantity and Billed Price refer t the detail characteristics Billing Date and Accunting Clerk. Scenari 2: Data Partitining by a Characteristic in an InfCube Dimensin (accunt data mdel) Examples One key figure f the data mdel is defined as InfObject, which is added t the fact table f the InfCube. The different semantic f this key figure is defined by a characteristic, which builds up a dimensin f the InfCube. The key figure Csts is part f the InfCube and is uniquely defined by the characteristics Value Type and Versin, which build up the dimensin Value Type/Versin f the InfCube. The characteristic Value Type can assume the values Plan, Actual, and s n. The key figure Ttal Sales Amunt is part f the InfCube and is uniquely defined by the characteristic Currency Type in a dimensin f the InfCube. The characteristic Currency Type can assume the values Lcal Currency, Dcument Currency, and s n. Mdeling MultiPrviders and InfSets with SAP BW.dc Page 4 14.06.2012

Scenari 3: Data Partitining by a MultiCube Example 1: Plan / Actual data in MultiCube Scenari Overview Queries: Plan / Actual Cmparisn Detailed View: Queries n Plan Data Basic InfCube 1 Plan Data Key Figure: Plan Csts MultiCube Plan / Actual Data Key Figures: Plan csts, Actual csts Detailed View: Queries n Actual Data Basic InfCube 2 Actual Data Key Figure: Actual Csts Hmgeneus data mdel Sme key figures (Plan Csts, Actual Csts) are defined as InfObjects. Each f them will be added t a different basic InfCube (Plan Data, Actual Data). Detailed Views (BW Queries) act n these basic InfCubes. Overview Queries use a MultiCube, which represents the verall reprting scenari (Plan / Actual Data Cmparisn). SAP AG 2003, Title f Presentatin, Speaker Name / 1 Mdeling MultiPrviders and InfSets with SAP BW.dc Page 5 14.06.2012

Scenari 3: Data Partitining by a MultiPrvider Example 2: Sales Order, Delivery, and Billing Data in a MultiPrvider Detailed View: Queries n Sales Order Data MultiPrvider Overview Data Key Figures: All Key Figures f the Basic InfCubes Characteristics: Custmer, Prduct, Sales Order Overview Queries: Custmer/Prduct/Sales Order Detailed View: Queries n Billing Data Basic InfCube 1 Sales Order Data Key Figures: Sales Order Quantity, Sales Order Price Characteristics: Custmer, Prduct, Sales Order, Sales Order Date, Sales Persn Basic InfCube 3 Billing Data Key Figures: Billed Quantity, Billed Price Characteristics: Custmer, Prduct, Sales Order, Billing Date, Accunting Clerk Basic InfCube 2 Delivery Data Key Figures: Delivered Quantity, Delivery Price Characteristics: Custmer, Prduct, Sales Order, Delivery Date, Delivery Persn Detailed View: Queries n Delivery Data Hetergeneus data mdel (acrss business scenari): Detailed data f the subscenaris (Sales Order, Delivery, Billing) are stred in the crrespnding basic InfCube. Detailed queries are defined with respect t the basic InfCubes Overview Queries use a MultiPrvider, which represents the verall reprting scenari (Reprts n Custmer/Prduct/Sales Order Level). SAP AG 2003, Title f Presentatin, Speaker Name / 1 Valuatin f the 3 scenaris The 3 scenaris described abve are cmpared in the fllwing table, where they are valuated with respect t their pssible uses cases. Mdeling MultiPrviders and InfSets with SAP BW.dc Page 6 14.06.2012

Data Partitining: Cmparisn f the Scenaris (1) Scenari 1: InfCube Cntains All Key Figures Scenari 2: Data Partitining by a Characteristic Scenari 3: Data Partitining by a MultiPrvider Use Cases fr each scenari Clsed Reprting Scenaris: Fixed number f key figures Hmgeneus data granularity All key figures are filled evenly in transactin data. Reprting Scenaris, which have t be extended easily: Additinal key figures crrespnd t additinal values f the partitining characteristic. Reprting scenari cnsists f several sub-scenaris: Fcus queries use the basic InfCubes f each subscenari. Overview queries, which require data frm several subscenaris use a MultiPrvider. Data Mdel Behavir, if the data granularity varies strngly. Characteristics and key figures in the tables f the InfCubes are distributed sparsely. Partly empty clumns exist This behavir may cause enhanced usage f data strage in the BW system. Parts f the data mdel crrespnding t different data granularity separated by the partitining characteristic. Additinal rws in fact table cmpared t scenari 1 N empty fields (i.e. key figures) exist in the fact table. Parts f the data mdel crrespnding t different data granularity separated by the definitin f basic InfCubes. MultiPrvider is nly a data view but nt a physical data strage. N redundancies ccur. Data Mdel Data Mdel derived frm the transactin data mdel in OLTP: All key figures defined as in the OLTP system cmplex and large fact table number f dimensins small cmpared t scenari 2 Lgical partining f the transactin data by using a suitable characteristic: less cmplex fact table, but higher number f dimensins cmpared t scenari 1 use partitining characteristic, if it exists in OLTP. Design f basic InfCubes reflects the sub-scenaris f the data mdel: less cmplex basic InfCubes smaller fact and dimensin tables Re-use f basic InfCubes in several MultiPrvider. SAP AG 2003, Title f Presentatin, Speaker Name / 1 Data Partitining: Cmparisn f the Scenaris (2) Scenari 1: InfCube Cntains All Key Figures Scenari 2: Data Partitining by a Characteristic Scenari 3: Data Partitining by a MultiPrvider Data Staging in BI Key figures can be transferred 1:1 frm OLTP transactinal data recrds int the InfCube. Split f transactinal data recrds in BW update rules: ne recrd with several key figures can be transfrmed int several recrds with ne key figure. calculate values f partitining characteristic Key figures can be transferred 1:1 frm OLTP transactinal data recrds int the basic InfCubes. Each sub-scenari has a separate InfSurce fr data lad parallel data staging is pssible Query Design Each key figure has clear semantics. The end-user can identify each key figure. Key figures have a clearly defined semantics nly in cntext with a fixed value f the partitining characteristic. The defintin and delivery f restricted key figures is necessary. Key figures in different basic InfCubes have t be distinct frm each ther. The definitin and delivery f key figures fr each subscenari (such as Plan Csts, Actual Csts) is necessary. Query Perfrmance Less cmplex data access t fact table cmpared t scenari 2 Empty and redundant fields may be selected. Mre cmplex selectin criteria cmpared t scenari 1 because f the use f restricted key figures in BW queries. Selected set f data is smaller cmpared t scenari 1. Fcus queries (such as nly Plan data) select n smaller data base cmpared t scenaris 1 and 2. Overview queries n MultiPrvider parallel data selectin fr each sub-scenari SAP AG 2003, Title f Presentatin, Speaker Name / 1 Mdeling MultiPrviders and InfSets with SAP BW.dc Page 7 14.06.2012

Recmmendatins If yu use scenari 3 (MultiPrvider), it is imprtant t mdel the sub-scenaris apprpriately. A split int mre than 10 sub-scenaris results in unclear data mdels. Therefre, yu can use up t 10 subscenaris t take different levels f data granularity int accunt. If yu want t read data nly frm a single basic InfPrvider, use the characteristic 0INFOPROV at query design t filter the basic InfPrvider in scenari 3 (MultiPrvider). Fr mre infrmatin, see SAP nte 728017. Often, a cmbinatin f the 3 different scenaris leads t a suitable data mdel in BI. If yu mdel cmpletely hmgenus data (such as ne basic InfPrvider fr each fiscal year that is a cnstant in these basic InfPrviders), use the database partitining functinality. Use Case: Data Mdel fr Plan / Actual Cmparisn with Plan Versins The reprting scenari cntains plan and actual data, which have a different granularity with respect t the time characteristics (Plan data per calendar / fiscal perid, actual data per calendar day). Plan data additinally refers t a characteristic Versin (fr example, defensive variant, and ffensive variant). Use a cmbinatin f scenari 3 and scenari 2. First, use scenari 3 t split the verall data mdel int tw basic InfCubes (Plan Data, Actual Data). Then, use scenari 2 fr the plan data: The basic InfCube Plan Data cntains ne characteristic in a dimensin Versin. Additinal versins laded frm the peratinal system are mapped by additinal values f this characteristic. Mdeling MultiPrviders and InfSets with SAP BW.dc Page 8 14.06.2012

3 Data Mdels Using InfSets InfSets are InfPrviders that lgically jin data and prvide this data fr BI queries. InfSets nly reference basic InfPrviders (InfCubes, DataStre bjects, master data InfObjects), but they cntain n data. All the BEx and OLAP services are available (authrizatins, texts, variables, hierarchies, calculated key figures) except navigatinal attributes f InfSet characteristics. In the InfSet maintenance, yu can make field descriptins unique fr the BEx user and hide fields f the basic InfPrviders that are nt imprtant fr reprting. When t use InfSets? T jin required data frm basic InfPrviders This allws building a relatinal BI data mdel with unified views fr reprting (several InfPrviders, but nly ne view). Therefre, we recmmend keeping data in smaller, basic InfPrviders that can be flexibly jined fr reprting purpses. Jin cncepts: T allw BEx Reprting n a DataStre bject withut turning the BEx Reprting indicatr n T evaluate time dependencies (fr example, jin time dependent master data InfObjects) T be able t create self jins and left uter jins Inner jin: A recrd can nly be in the selected result set if there are entries in bth jined tables Left uter jin: If there is n crrespnding recrd in the right table, the recrd is part f the result set (fields belnging t the right table have initial values) Tempral jin: A jin is called tempral if at least ne member is time-dependent. Self jin: The same bject is jined tgether 3.1 Secnd-Order Navigatinal Attributes Use case An InfPrvider cntains characteristics that have navigatinal attributes. These are the first-rder navigatinal attributes f the InfPrvider. These first rder navigatinal attributes again have navigatinal attributes. These are the secndrder navigatinal attributes f the InfPrvider. The end-user wants t navigate / drill-dwn t these secnd-rder navigatinal attributes in BI queries based n the InfPrvider. Mdeling MultiPrviders and InfSets with SAP BW.dc Page 9 14.06.2012

Example: Transactin Data Using Cnslidated InfObjects 1. DataStre bject Purchase Order (ZPUR_O01) cntains cnslidated InfObjects 0PRODUCT and 0GN_VENDOR. 2. Cnslidated InfObjects 0PRODUCT and 0GN_VENDOR cntain lcal-view InfObjects 0MATERIAL (R/3) / 0BBP_PROD (SRM) and 0VENDOR (R/3) / 0BPARTNER (SRM) as first-rder navigatinal attributes. 3. Lcal-View InfObjects 0MATERIAL (R/3) / 0BBP_PROD (SRM) and 0VENDOR (R/3) / 0BPARTNER (SRM) cntain lcal-view attributes as secndrder navigatinal attributes with respect t the DataStre bject Purchase Order (ZPUR_O01). POGUID 0PRODUCT 0GN_VENDOR... QUANTITY VALUE 0PRODUCT... 0MATERIAL 0BBP_PROD... 0GN_VENDOR... 0VENDOR 0BPARTNER... 0MATERIAL... 0MATL_GROUP... 0BBP_PROD... PROD_TYPE... 0VENDOR... 0INDUSTRY... 0BPARTNER... 0IND_CODE... 4. In queries based n DataStre bject Purchase Order (ZPUR_O01), the user wants t navigate / drill-dwn t the secnd-rder navigatinal attributes 0MATL_GROUP, 0PROD_TYPE, 0INDUSTRY, 0IND_CODE. SAP AG 2004, Title f Presentatin / Speaker Name / 1 Excursin: Why d cnslidated InfObjects need secnd-rder navigatinal attributes? One basic idea f the cnslidated InfObjects is t keep these InfObjects as lean as pssible; that is, nly attributes used fr glbal (crss surce systems and applicatins) reprting are relevant. The cnslidated InfObject cannt cntain any attribute f the lcal views. Glbal attributes delivered by SAP can nly be: Cnslidated InfObjects InfObjects with unified r standardized values in all surce systems DUNS number, ISO Cdes, UNSPSC Cdes, and s n. Mdeling MultiPrviders and InfSets with SAP BW.dc Page 10 14.06.2012

Slutin 1: Use InfSet t Build a Star Schema Create an InfSet starting with the InfPrvider. Jin the master data tables f the characteristics, which lead t the secnd-rder navigatinal attributes, t the InfPrvider. Jin the master data tables f the first-rder navigatinal attributes, which cntain the secnd-rder navigatinal attributes, t the master data tables f the characteristics. Define queries based n the InfSet using the secnd-rder navigatinal attributes fr navigatin / drill dwn. Example: 0MATERIAL... 0MATL_GROUP... 0BBP_PROD... PROD_TYPE... 0PRODUCT... 0MATERIAL 0BBP_PROD... DataStre Object ZPUR_O01 POGUID 0PRODUCT 0GN_VENDOR... QUANTITY VALUE Inner jin 0GN_VENDOR... 0VENDOR 0BPARTNER... 0VENDOR... 0INDUSTRY... 0BPARTNER... 0IND_CODE... SAP AG 2004, Title f Presentatin / Speaker Name / 1 Arguments pr and cntra this slutin (the use f an InfSet t mdel secnd-rder navigatinal attributes) Pr: Cntra: All cmpnents f the data mdel (InfPrvider, InfObjects) can be kept lean. There are n data redundancies. N additinal master data attributes have t be determined during extractin r uplad f transactin data. Clear, intuitive, and flexible data mdel: Yu can add additinal secnd-rder navigatinal attributes can by extensin f the InfSet. Therefre, n additinal data lad is required after a data mdel extensin. Perfrmance prblems may ccur at BI query runtime if t many master data table f first- and secnd-rder navigatinal attributes are jined t the InfPrvider. There are ther slutins withut the use f InfSets t mdel secnd-rder navigatinal attributes. Slutin 2 (Extensin f the InfPrvider): Add the first-rder navigatinal attributes, which cntain the secnd-rder navigatinal attributes, t the characteristics f the InfPrvider (in additin t the existing characteristics). Slutin 3 (Extensin f the characteristics): Add the secnd-rder navigatinal attributes, as navigatinal attributes t the characteristics f the InfPrvider. Mdeling MultiPrviders and InfSets with SAP BW.dc Page 11 14.06.2012

Slutins 2 and 3 have better perfrmance at BI query runtime, if the InfPrvider is an InfCube: The relatinship between first- and secnd-rder navigatinal attributes is fixed by the BI data mdel and data is persistently laded t InfPrviders (slutin 2) r master data tables (slutin 3). The BI data mdel f slutins 2 and 3 cause redundant data in BI inflexible data mdel: Additinal characteristics (such as first-rder navigatinal attributes) in the InfPrvider cause a relad f transactinal data realignment prblems within the InfPrvider if the assignment f InfObjects t first-rder navigatinal attributes changes Fr each use case, where secnd-rder navigatinal attributes are needed, determine if the BI data mdel has t be flexible with regard t changes (slutin 1) r if the BI query perfrmance has t be ptimized (slutin 2 r 3). 3.2 Data Mdel Examples fr Tempral Jins Example 1: Tw time-dependent master data InfObjects are jined. The InfObject, Cst Center (technical name CSTCNTR) is jined t a secnd InfObject, Prfit Center (technical name PROFITC). InfObject, Cst Center, has the attribute Prfit Center, which is used as jin cnditin. Bth InfObjects are time-dependent and cntain the attribute, Respnsible Persn (technical name RESPPERS). The tempral jin gives insight t the questin: Which persn respnsible fr a cst center has wrked tgether with a certain persn respnsible fr a prfit center at the same time? Master data table f InfObject Cst Center: CSTCNTR (key) DATETO (key) DATEFROM PROFITC RESPPERS 4711 31.05.2001 BI Jack 4711 31.12.2001 BI Jhn 4711 BI Je 0815 31.01.2001 KM Jane 0815 01.02.2001 KM Jill Master data table f InfObject Prfit Center: PROFITC (key) DATETO (key) DATEFROM RESPPERS BI 30.06.2001 01.01.2000 Karl BI 01.07.2001 Hanna The InfSet jins bth tables using the cnditin: PROFITC f master data table CSTCNTR is equal t the PROFITC f master data table PROFITC. A query uses this InfSet and selects Cst Center 4711. Withut cnsideratin f any time-dependencies, the result f the query includes 6 recrds. Fr simplicity, the clumns CSTCNTR = 4711 (query selectin) and PROFITC = BI (jin cnditin) are left ut: RESPPERS (CSTCNTR) RESPPERS (PROFITC) DATEFROM (CSTCNTR) DATETO (CSTCNTR) DATEFROM (PROFITC) DATETO (PROFITC) Jack Karl 31.05.2001 01.01.2000 30.06.2001 Jhn Karl 31.12.2001 01.01.2000 30.06.2001 Mdeling MultiPrviders and InfSets with SAP BW.dc Page 12 14.06.2012

Je Karl 01.01.2000 30.06.2001 Jack Hanna 31.05.2001 01.07.2001 Jhn Hanna 31.12.2001 01.07.2001 Je Hanna 01.07.2001 The result in cnsideratin f time-dependencies (nly 4 valid recrds) is shwn in the fllwing figure. Result in Cnsideratin f Time Dependencies RESPPERS RESPPERS [PROFITC] DATEFROM DATETO DATEFROM [PROFITC] DATETO [PROFITC] Jack Karl 31.05.2001 01.01.2000 30.06.2001 Jhn Karl 31.12.2001 01.01.2000 30.06.2001 Je Karl 01.01.2000 30.06.2001 Jack Hanna 31.05.2001 01.07.2001 Jhn Hanna 31.12.2001 01.07.2001 Je Hanna 01.07.2001 Karl Hanna PROFITC = BW CSTCTR = 4711 Jack Jhn Je 01.07.2001 tday t InfSets - Christel Rüger The tw recrds marked in red d nt have verlapping time intervals [DATEFROM; DATETO] fr Cst Center and Prfit Center master data tables. Therefre, these recrds are nt selected by the tempral jin. The ther fur recrds have an verlapping time interval and are returned t the BI query. The bundaries DATEFROM and DATETO f valid time interval fr each recrd are marked by green fields and are als displayed in a time bar diagram belw. Example 2: Time-dependent master data InfObject is jined t a DataStre bject cntaining a key date. The InfObject, Cst Center (technical name CSTCNTR) is jined t a DataStre bject, Invice (technical name INVOICE). The DataStre bject has the characteristic Cst Center, which is used as a jin cnditin. The InfObject CSTCNTR is time dependent, whereas the DataStre bject INVOICE has the InfObject calendar day (technical name 0CALDAY), which can be used as key date. The tempral jin gives insight t the questin: Wh was the respnsible persn f a Cst Center at the date when the invice was psted? Mdeling MultiPrviders and InfSets with SAP BW.dc Page 13 14.06.2012

Master data table f InfObject Cst Center: CSTCNTR (key) DATETO (key) DATEFROM RESPPERS 4711 31.05.2001 Jack 4711 31.12.2001 Jhn 4711 Je 0815 31.01.2001 Jane 0815 01.02.2001 Jill DataStre Object INVOICE: INVOICE ID (key) 0CALDAY CSTCNTR 100001 03.02.2001 4711 100002 28.04.2002 4711 The InfSet jins bth tables using the cnditin: CSTCNTR f master data table CSTCNTR is equal t the CSTCNTR f DataStre Object INVOICE. The result in cnsideratin f time-dependencies (nly 2 valid recrds) is shwn in the fllwing figure. Result in Cnsideratin f Time Dependencies RESPPERS INVOICE ID [INVOICE] DATEFROM DATETO 0CALDAY [INVOICE] Jack 100001 31.05.2001 03.02.2001 Jhn 100001 31.12.2001 03.02.2001 Je 100001 03.02.2001 Jack 100002 31.05.2001 28.04.2002 Jhn 100002 31.12.2001 28.04.2002 Je 100002 28.04.2002 100001 CSTCNTR = 4711 100002 CSTCNTR = 4711 Jack Jhn Je 03.02.2001 28.04.2002 tday t InfSets - Christel Rüger Mdeling MultiPrviders and InfSets with SAP BW.dc Page 14 14.06.2012

A recrd f the InfSet is selected by the BI query nly if the key date f the DataStre bject INVOICE is within the valid time interval [DATEFROM; DATETO] f the master data InfObject CSTCNTR. The derivatin f a key date is als pssible fr the time characteristics f the SAP BI system: 0CALWEEK Calendar Year / Week 0CALMONTH Calendar Year / Mnth 0CALQUARTER Calendar Year / Quarter 0CALYEAR Calendar Year 0FISCPER Fiscal Year / Perid (cmpund with fiscal year variant 0FISCVARNT) 0FISCYEAR Fiscal Year (cmpund with the fiscal year variant 0FISCVARNT) Yu can chse the begin date r end date f the abve time intervals r a fixed date within these intervals. Example 3: Time-dependent master data InfObject is jined t a DataStre bject, where a valid time interval can be derived frm a time characteristic. The InfObject, Cst Center (technical name CSTCNTR) is jined t a DataStre bject, Prject Spnsr (technical name SPONSOR). The DataStre bject has the characteristic Cst Center, which is used as jin cnditin. The InfObject CSTCNTR is time dependent, whereas the DataStre bject SPONSOR has the key field Calendar Mnth (technical name 0CALMONTH) t define the valid time interval f each recrd. The tempral jin gives insight t the questin: Wh was the respnsible persn f a Cst Center in the time interval when a prject f the Cst Center was spnsred? Master data table f InfObject Cst Center: CSTCNTR (key) DATETO (key) DATEFROM RESPPERS 4711 31.05.2001 Jack 4711 31.12.2001 Jhn 4711 Je 0815 31.01.2001 Jane 0815 01.02.2001 Jill DataStre Object SPONSOR: SPONSOR (key) 0CALMONTH (key) CSTCNTR Hug 04.2003 4711 Alice 02.2001 4711 The InfSet jins bth tables using the cnditin: CSTCNTR f master data table CSTCNTR is equal t the CSTCNTR f DataStre Object SPONSOR. The time interval [DATEFROM; DATETO] fr each recrd in the DataStre Object SPONSOR can nw be derived frm the time characteristic 0CALMONTH: 0CALMONTH (Surce) DATEFROM (Target) DATETO (Target) 02.2001 01.02.2001 28.02.2001 04.2003 01.04.2003 30.04.2003 Mdeling MultiPrviders and InfSets with SAP BW.dc Page 15 14.06.2012

Similarly, yu can retrieve the valid time interval [DATEFROM; DATETO] fr the time characteristics f the SAP BI system: 0CALWEEK Calendar Year / Week 0CALMONTH Calendar Year / Mnth 0CALQUARTER Calendar Year / Quarter 0CALYEAR Calendar Year 0FISCPER Fiscal Year / Perid (cmpund with the fiscal year variant 0FISCVARNT) 0FISCYEAR Fiscal Year (cmpund with the fiscal year variant 0FISCVARNT) The result in cnsideratin f time-dependencies (nly 2 valid recrds) is shwn in the fllwing figure. Result in Cnsideratin f Time Dependencies RESPPERS SPONSOR [SPONSOR] DATEFROM DATETO DATEFROM [SPONSOR] DATETO [SPONSOR] Jack Alice 31.05.2001 01.02.2001 28.02.2001 Jhn Alice 31.12.2001 01.02.2001 28.02.2001 Je Alice 01.02.2001 28.02.2001 Jack Hug 31.05.2001 01.04.2003 30.04.2003 Jhn Hug 31.12.2001 01.04.2003 30.04.2003 Je Hug 01.04.2003 30.04.2003 Alice Hug CSTCNTR = 4711 Jack Jhn Je CSTCNTR = 4711 01.02.2001 28.02.2001 01.04.2003 30.04.2003 tday t InfSets - Christel Rüger The fur recrds marked in red d nt have verlapping time intervals [DATEFROM; DATETO] fr Cst Center master data tables and DataStre Object Spnsr. The ther tw recrds have an verlapping time interval and are returned t the query. The bundaries DATEFROM and DATETO f valid time interval fr each recrd are marked by green fields and are als displayed in a time bar diagram. Mdeling MultiPrviders and InfSets with SAP BW.dc Page 16 14.06.2012

4 Perfrmance Aspects f MultiPrviders and InfSets MultiPrviders InfSets These are the advantages f MultiPrviders: Lcal queries (n each basic InfPrvider) versus glbal queries (n MultiPrvider, parallel executin). Independent (and parallel) data lad int the basic InfPrviders. Small ttal data vlumes: basic InfPrvider have less redundant data; they are sparsely filled and less cmplex. As a general rule, we recmmend assigning up t 10 basic InfPrviders t a MultiPrvider. If the number f basic InfPrviders is significantly higher, the verhead in cmbining results at BI query runtime may becme excessive. If a MultiPrvider cntains at least ne basic InfPrvider with nn-cumulative key figures, all queries are prcessed sequentially. The perfrmance ptimizing tls f the OLAP (such as caching, aggregatin) nly wrk fr a MultiPrvider if all cnstituent basic InfPrviders f the MultiPrvider supprt these tls. InfSets d nt have the set f perfrmance tls as InfCubes (such as aggregates, partitining, and cmpressin). Use left uter jins in InfSets nly when necessary. A left uter jin has a significantly wrse perfrmance than a crrespnding inner jin. If yur reprting requirements n a DataStre Object are very restricted (that is, yu want t display nly very few, selective recrds), use an InfSet n tp f the DataStre bject and disable the BEx Reprting indicatr. This results in better data lading perfrmance, but als in wrse perfrmance at BI query runtime if mre than 10 recrds are selected frm the DataStre Object. 5 List f Dcuments Related t MultiPrviders and InfSets MultiPrviders: Fr mre infrmatin, see the SDN dcument Hw t Create Efficient MultiPrvider Queries https://www.sdn.sap.cm/irj/servlet/prt/prtal/prtrt/dcs/library/uuid/751be690-0201-0010-5e80- f4f92fb4e4ab InfSets: Fr mre infrmatin, see fllwing OSS ntes: 583249: Tempral jins 592785: Interpretatin f results 577953: Left uter jins Mdeling MultiPrviders and InfSets with SAP BW.dc Page 17 14.06.2012