You are on page 1of 16

BW Training Topics

1 Data Modeling
2 Extended Star Schema & Aggregates

3 InfoObjects, DataSources & InfoSources

4 BW overview & Administrative Workbench

5 Data Staging, Data Storage & ODS

6 Data Extraction

7 Reporting & Analysis and Tracking History

8 Business Explorer

9 Query & Analysis


Staging Engine
Staging Engine consists of several processes that gather
data from data sources, clean and perform data
transformation and then populate the InfoCubes.
The staging process is described in the following 5 steps:
1. SAP R/3 prepares needed data using the Extract
Structure.
2. The Extractor then moves data to the Transfer
Structure in the R/3.
3. Using ALE or tRFC, R/3 copies data to BW in the
form of Transfer Structure.
4. Using Tfr. Rules in BW the data is transferred into
the Communication Structure.
5. The Update Rules update the InfoCube.
Architecture - PSA

Persistent Staging Area (PSA)


 Based upon transfer structure

 Source system dependent

 Transaction data

 Master data

 Inbound layer
PSA
PSA

DataSource DataSource

Data extract Data extract


OLTP System
OLTP System
PSA (Persistent Staging Area)

Data in the PSA is based upon the transfer structure


and is not transformed via transfer rules. The data is all
system dependent. Data in the PSA can be transaction
data as well as master data.
The PSA is a transparent table for storing the detailed
requests in the format of the transfer structure. The
PSA is defined according to DataSource and source
system.
The update of the PSA and the InfoCube can run
sequentially or in parallel.
ODS – Business Scenario

Sometimes it is desirable to combine data from


different DataSources before the same is stored
into the InfoCubes.
Also, there are analyses that need access to the
detailed data than that found in the Cubes.
Motivation for ODS

Consolidation and Cleansing

Store data on document level

Overwrite capability of characteristics

Reporting

 Direct on this document level data

 Drilldown from InfoCube to document level


Motivation for ODS ...

1. Not possible to overwrite data in the InfoCube as


whenever data is added to the InfoCube, this data is
aggregated.

2. Data can be overwritten in the ODS and this provides a


significant capability to BW.

3. A further motivation is the need for a place where the data


can be consolidated and cleansed. This is important when
we upload data from completely different Source
Systems.

4. After consolidation and cleansing, the data can be


uploaded to the InfoCubes.
ODS...

1. Prior to the existence of the ODS, decisions on granularity


were based solely on the data in the InfoCube.

2. Now the InfoCube can be less granular with data held for
a longer period of time versus the ODS which can be very
granular but hold data for a shorter period of time.
BW PSA/ODS Architecture
B
E
x
OLAP Processor

B
W
Update Rules Update Rules Update Rules Update Rules

BW Operational Data Store


S
Transfer Rules Transfer Rules E
R
PSA PSA V
E
DataSource A DataSource B R
Architecture – BW ODS

BW Operational Data Store (ODS)

 Merge data from different InfoSources

 Transaction data

 Technical and data changes

BW ODS
ODS
ODSObjects
Objects
 Denormalized data structure -
Transparent table
Update Rules
 Data available for reporting Transfer Rules

 Potential to create
an external file
PSA
PSA
ODS 2.0B - Data Flow
Communication Structure Update
Update
Rules
Rules

ODS InfoCube

Update
Update
No changes
Rules
Rules possible
Corresponds
to the ODS
Communication Structure structure
Transfer
Transfer
Rules
Rules

Transfer Structure

PSA

Transfer Structure

OLTP
OLTP System
System
BW ODS Object: Update Example

BEx ODS
ODS-
ODS- PSA
PSA
Reports Object
Object
Data Mart
Active Data Change Log
ODS Object

123 old
123 old/
new
Doc No. Item Amount Status
123 1 100 Open Activate
123 1 60 Delivered

Doc No. Item Amount Status


123 new
123 1 -100 Open
Doc No. Item Amount Status 123 1 60 Delivered
123 1 60 Delivered Modified +
New Data

Notes:
1. Entire ODS key of request ID, artificial record ID and semantic ID (ex. order
number, item number, sched. line) not shown above.
2. The records in italics will be in the tables after activation.

 SAP AG 1999
ODS - Reporting
1 2 3

ODS Only InfoCube to ODS ODS to OLTP

RRI
OLAP Processor OLAP Processor OLAP Processor OLTP System

ODS ODS
Object InfoCube Object
Report Report Interface (RRI)

Here one query analyzes the data of an InfoCube and then


uses the RRI to drill down to the ODS to get more detail.

Interface to start a receiver report with selections from


a sender report

BEx query is the sender

Possible receivers (jump targets) are

 BEx query

 R/3 transaction and report

 InfoSet Query (formerly SAP Query)


ODS 2.0B - Data Flow Scenario

InfoCube

Update Rules

Order-Deliv.
ODS
Order Delivery

Update Rules
Transfer Rules

PSA
PSA
Order Delivery

DataSource DataSource

OLTP System Data extract Data extract


Data extract Data extract
ODS 2.0B - Summary
Rename ODS into Persistent Staging Area (PSA)
 Inbound layer of the BW

ODS-Object as a transparent table


 Physical key, e.g. document no., item w/o technical

fields like request id


 Update of non-keyfields supported

Enhancement of current staging architecture


Consolidated view of the data
 Based on the communication structure

Update of InfoCubes from ODS-Objects


 Automatic delta computation

Fully integrated into Administrator Workbench


Reporting via Business Explorer
Enabled for BW Data Mart Interface

 SAP AG 2000

You might also like