BW Quiz From SearchSAP
BW Quiz From SearchSAP
BW Quiz From SearchSAP
Note: This quiz is based on the BW release 3.5, which is a part of SAP NetWeaver 2004.
Data warehousing is the foundation for all business intelligence (BI) solutions, and SAP
BW is no exception. If you're an aspiring SAP BI/BW consultant, it is important to have
extensive knowledge and experience with SAP BW.
SearchSAP.com's quiz series is designed to help you evaluate your BI/BW skill set. This
quiz is the final part of the three-part series written by SAP BI/BW expert Jay Narayanan.
It is an advanced-level quiz and it covers Multiproviders and Infosets, change runs,
navigational attributes and other advanced BW scenerios.
NOTE: Some questions may have more than one appropriate answer!
a. RSA3
b. RSA4
c. RSA5
d. RSA6
ANSWER(S): A, C, D
Transaction RSA3, or extractor checker, is used in the BW source system to check data
sources for various extraction modes, including full update, delta update and delta
initialization.
RSA5 is for installing standard business content data sources and RSA6 is for
maintaining data sources.
2. True or False? A reference characteristic will use the sid table and master data
table of the referred characteristic.
a. True
b. False
ANSWER(S): A
ANSWER(S): D
All the statements except D are true. It is possible to change a navigational attribute back
to a display attribute in an info cube, without deleting all data from the info cube.
4. True or False? It is possible to create a key figure without assigning currency or
unit.
a. True
b. False
ANSWER(S): A
Yes, it is possible to create a key figure without assigning a unit if the data type is one of
these four: Number, Integer, Date or Time.
a. An info cube needs to contain all info objects of the compounded info object if it has
been included in the info cube.
b. An info object cannot be included as a compounding object if it is defined as an
attribute only.
c. An info object can be included as an attribute and a compounding object
simultaneously.
d. The total length of a compounded info object cannot exceed 60.
ANSWER(S): A, B, D
When a compounded info object is included in an info cube, all corresponding info
objects are added to the info cube. If an info object is defined as an attribute, it cannot be
included as compounding object. The total length of the compounding info objects cannot
exceed 60 characters.
ANSWER(S): A, C
Any characteristic in the info cube should be assigned to a dimension. One characteristic
cannot be assigned to more than one dimension. One dimension can have more than one
characteristic, provided it is not defined as a line item dimension.
7. The following statements are true for info cubes and aggregates.
ANSWER(S): A, C
Once the info cubes are compressed it is not possible to delete data based on the
requests. There won't be request IDs anymore. Requests can be deleted even if the
requests in aggregates have been compressed. But the aggregates will have to be de-
activated. Deleting an uncompressed request from an info cube will automatically delete
the corresponding request from aggregate if the aggregate request has not been
compressed.
8. The following statements are true regarding the ODS request deletion.
a. It is not possible to delete a request from ODS after the request has been activated.
b. Deleting an (inactive) request will delete all requests that have been loaded into the
ODS after this request was loaded.
c. Deleting an active request will delete the request from the change log table.
d. None of the above.
ANSWER(S): C
It is possible to delete requests from an ODS, even if the request has been activated. The
"before and after image" of the data will be stored in the change log table using which
the request will be deleted.
Deleting a request which has not been activated in ODS will not delete the requests
which are loaded after this request. But if the request has been activated then the loaded
and activated requests later will get deleted. Also the change log entries will be deleted
for that request.
ANSWER(S): A, B, C
Aggregates summarize and store data from an info cube. Characteristics and
navigational attributes of an info cube can be used to create aggregates. Since
aggregates contain summarized data, the amount of data in aggregates will be much less
that the cube which makes the queries to run faster when they access aggregates.
10. True or False? If an info cube has active aggregates built on it, the new requests
loaded will not be available for reporting until the rollup has been completed
successfully.
a. True
b. False
ANSWER(S): A
Newly-loaded requests in an info cube with aggregates will not be available for reporting
until the aggregate rollup has been completed successfully. This is to make sure that the
cube and aggregates are consistent while reporting.
a. To deliver structured information that the business user can easily navigate by using
any possible combination of business terms to show the KPIs.
b. To make it easier for developers to build applications, that will be helpful for the
business users.
c. To make it easier to store data in the database and avoid redundancy.
d. All of the above.
ANSWER(S): A
a. If a cube has been partitioned, the E table of the info cube will be partitioned on time.
b. The F table of the info cube is partitioned on request.
c. The PSA tableis partitioned automatically with several requests on one partition.
d. It is not possible to partition the info cube after data has been loaded, unless all the data
is deleted from the cube.
ANSWER(S): A, B, C, D, F
BW allows partitioning of the info cubes based on time. If the info cube is partitioned, the
e-fact table of the info cube will be partitioned on the time characteristic selected.
The F fact table is partitioned on request ids automatically during the loads. PSA tables
are also partitioned during the loads and can accommodate more than one request. For
an info cube to be partitioned, all data needs to be removed from the info cube.
a. Query navigation states and query results are stored in the application server memory.
b. If the same query has been executed by another user the result sets can be used if the
global cache is active.
c. Reading query results from OLAP Cache is faster than reading from the database.
d. Changing the query will invalidate the OLAP Cache for that query.
ANSWER(S): A, B, C, D
Query results are stored in the memory of application server, which can be retrieved later
by another user running the same query. This will make the query faster since the results
are already calculated and stored in the memory. By changing the query, the OLAP
Cache gets invalidated.
14. The following statements are true about the communication structure.
ANSWER(S): A, B
The communication structure contains all info objects in the info source and it is used to
update the info cube by temporarily storing the data that needs to be updates to the data
target. It doesn't depend on the source system.
ANSWER(S): B, D
An ODS cannot be created without any data fields, and it can have a maximum of only 16
key fields. Key figures cannot be included as a key field in an ODS. The export data
source is created after an ODS has been created and activated.
BW Quiz – Level 2: Intermediate
1. Identify the statement(s) that is/are true. A change run...
ANSWER(S): A, B
Change run activates the Master data and Hierarchy data changes. Before the activation
of these changes, all the aggregates that are affected by these changes are realigned.
Realignment is not necessarily done by reading InfoCubes. If these are part of another
aggregate that can be used to read data for the realignment, change run uses that
aggregate.
ANSWER(S): A, B
Multiproviders are like virtual Infoproviders that do not store any data. Basic InfoCubes,
ODSs, info sets or Info objects can be used to build a Multiprovider. Multiproviders can
even be built on a single Infoprovider.
3. The structure of the PSA table created for an info source will be...
ANSWER(S): D
The structure of PSA tables will have an initial four fields: request id, packet number,
partition value and record number. The remaining fields will be exactly like Transfer
Structure.
4. In BW, special characters are not permitted unless it has been defined using this
transaction:
a. rrmx
b. rskc
c. rsa15
d. rrbs
ANSWER(S): B
a. One info source can have more than one source systems assigned to it
b. One info source can have more than one data source assigned to it provided the data
sources are in different source systems
c. Communication structures is a part of an info source
d. None of the above
ANSWER(S): A, C
Info sources can be assigned to multiple source systems. Also, info sources can have
multiple data sources within the same source system. Communication structure is a part
of the source system.
6. Select the statement(s) that is/are true about the data sources in a BW system:
a. If the hide field indicator is set in a data source, this field will not be transferred to BW
even after replicating the data source
b. A field in a data source won't be usable unless the selection field indicator has been set
in the data source
c. A field in an info package will not be visible for filtering unless the selection field has
been checked in the data source
d. All of the above
ANSWER(S): A, C
If the hide field is checked in a data source, that field will not be transferred to a BW
system from the source system even after replication. If the selection field is not checked,
that field won't be available for filtering the info package.
7. Select the statement(s) which is/are true about the 'Control parameters for data
transfer from the Source System'
ANSWER(S): A
ROIDOCPRMS is the table in the BW source system that is used to store the parameters
for transferring data to BW. Max size is the size in KB which is used to calculate the
number of records in each packet. Max lines is the maximum number of records in each
packet.
8. The indicator 'Do not condense requests into one request when activation takes
place' during ODS activation applies to condensation of multiple requests into one
request to store it in the active table of the ODS.
a. True
b. False
ANSWER(S): B
This indicator is used to make sure that the change log data is not compressed to one
request when activating multiple requests at the same time. If these requests are
combined to one request in change log table, individual deletion will not be possible.
9. Select the statement(s) which is/are not true related to flat file uploads:
ANSWER(S): C
ANSWER(S): A, C
Dimensional attributes have a performance advantage while running queries since the
number of table joins will be less compared to navigational attributes. For navigational
attributes, the history of changes will not be available. But for dimensional attributes, the
InfoCube will have the change history.
11. When a Dimension is created as a line item dimension in a cube, Dimensions IDs
will be same as that of SIDs.
a. True
b. False
ANSWER(S): A
When a Dimension is created as a line item dimension, the SIDs of the characteristic is
directly stored in the fact tables and these are used as Dimension IDs. Dimension table
will be a view off of SID table and fact table.
12. Select the true statement(s) related to the start routine in the update rules:
ANSWER(S): A, B, C
In the start routine, all records are available for processing. Variables declared in the
global area can be used in individual routines. Returncode greater than 0 will abort
processing of all records in the packet.
a. True
b. False
ANSWER(S): A
14. After any changes have been done to an info set it needs to be adjusted using
transaction RSISET.
a. True
b. False
ANSWER(S): A
After makeing any type of change to an info set, it needs to be adjusted using the
transaction RSISET.
a. Read mode determines how the OLAP processor retrieves data during query execution
and navigation
b. Three different types of read modes are available
c. Can be set only at individual query level
d. None of the above
ANSWER(S): A, B
Read mode determines how an OLAP processor retrieves data during query execution
and navigation. Three types of read modes are available:
1. Read data during expand hierarchies
2. Read data during navigation
3. Read data all at once
Read mode can be set at info provider level and query level.
BW Quiz – Level 3: Advanced
1. Select the correct statements about the steps executed by a change run.
a. The steps activate the new master data and hierarchy data changes.
b. All aggregates are realigned and recalculated.
c. Aggregates containing navigational attributes are realigned and recalculated for the
master data changes.
d. The steps delete the 'A' (active) records for which the 'M' (modified) records exist from
master data tables, and makes all modified records active.
e. All of the above.
ANSWERS: A, C, D
Master data and hierarchy data changes are activated and all the aggregates which have
navigational attributes which will be affected by the changes are realigned. Change run
deletes all active records for which modified records exists in master data 'P' table and
makes all modified records to active.
2. Key figures that are set for exception aggregation MIN or MAX in an aggregate
cause the aggregates to be completely rebuilt for each change run alignment.
a. True
b. False
ANSWER: A
If an aggregate contains key figures which are built as MIN or MAX, that will force a
change run to recreate these aggregates during the alignment process.
ANSWERS: B, D
Unless specified using transaction RSKC, special characters cannot be loaded into BW
fields where it needs to generte SIDs. So it is possible for these characters to be loaded
into text fields.
4. A change run updates the 'E' table of the aggregates while doing the alignment for
changes in the master data.
a. True
b. False
ANSWER: B
A change run doesn't update the 'E' table of the aggregates for the alignment of
aggregates due to master data changes. The alignment is done by inserting rows with the
necessary negative and positive key figure values in the 'F' table.
5. Select the correct statements related to the control parameters for a data transfer
in table ROIDOCPRMS in the BW source system.
a. The field MAXSIZE is the maximum number of records which can be transferred to
BW in a single packet.
b. The field MAXSIZE is the size in KB which is used to calculate the number of records
per data packet.
c. MAXLINES is the maximum number of records which can be transferred to BW per
data load.
d. If the number of data records per packet exceeds MAXLINES value the extraction
process fails.
ANSWER: B
In the table ROIDOCPRMS, MAXSIZE is the size in kilobytes which is used to calculate
the number of records in each datapacket to be transferred to BW. If the calculated
number of records exceeds MAXLINES, the packet size in terms of number of records will
be made the same as the value of MAXLINES value.
a. Both Multiproviders and Infosets can contain all the info providers in BW.
b. Queries built on Multiproviders use 'union' and queries on Infosets use 'join' to retrieve
data from different info providers.
c. Both Multiproviders and Infosets do not have data, but data is accessed from the basic
info providers used in these objects.
d. None of the above.
ANSWERS: B, C
ANSWERS: A, C, D
Transaction RSRCACHE is for the Cache Monitor. The OLAP Cache will be active
unless the 'cache inactive' flag is set. Persistent mode is to specify the action to be taken
when the cache memory is exhausted.
a. Performance of the ODS activation improves when the BEx reporting flag is switched
to off.
b. Overwriting a data record is not allowed if the 'unique' data record flag is set.
c. Data targets are updated from the ODS regardless of the ODS activation status.
d. All of the above.
ANSWERS: A, B
If the BEx Reporting flag is switched off than the SID won't have to be taken when
activating the ODS. This improves performance. If the 'unique data records' flag is set, it
is not possible to load records to the ODS for which the key combination already exists.
Data targets can be updated only after the ODS activation takes place successfully.
9. It is not possible to activate an ODS which contains a request from a full load and
a Delta Initialization load of the same data source.
a. True
b. False
ANSWER: B
It is possible to activate an ODS which contains Delta and full loads from the same
InfoSource if the full load is done with the repair flag set in the InfoPackage.
10. Select the correct statements regarding data deletion settings in an InfoPackage.
a. It is possible to set an InfoPackage to delete all the data in an InfoCube during the
loads.
b. Only uncompressed data can be set to be deleted from the cube in an InfoPackage
during the loads.
c. Deletion settings can be done only for basic InfoCubes.
d. Data deletion settings in an InfoPackage are possible only for full loads.
e. All of the above.
ANSWERS: A, C
All data can be set to be deleted from a basic cube during full loads and Delta loads.
Data can be deleted depending on various conditions. Data can only be deleted from
basic cubes, not the ODS.
a. Multiprovider queries create one process per info provider involved and are processed
parallel by default.
b. It is not possible to make Multiprovider queries run sequential.
c. Multiprovider queries create a parent process which provides a synchronization point
to collect the overall result from other sub processes.
d. Parallel processing is always faster than sequential processing in Multiproviders.
e. All of the above.
ANSWERS: A, C
Multiprovider queries create processes to run on the individual info providers involved.
These processes run parallel by default. The parent process provides a synchronization
point to collect overall results from sub processes. Many times the parallel processing
may be slower than sequential processing, if the data volume is high.
12. Select the correct statements about the ALPHA conversion routine in BW.
ANSWERS: A, D, E
The ALPHA conversion routine is assigned to the characteristic info objects when
created. This needs to be deleted manually if not required. The conversion is applied to
'external to internal' formats and vice versa. For 'external to internal' formats, the values
are right aligned and the spaces on left side is padded with zeros.
ANSWERS: A, B, D
Queries using navigational attributes will be slower since additional tables have to be
joined with the Fact tables to get the desired results. If there are changes to the values of
navigational attributes, the aggregates using these will have to be readjusted, which is
done by a change run. Navigational attributes can be made to display only without
removing the data from the cube.
14. A 'Check for Referential Integrity' can only be possible for information sources
with flexible updating.
a. True
b. False
ANSWER: A
A referential integrity check is only possible for info sources with flexible updating.
a. New partitions on F table of an InfoCube are created during data loads to the InfoCube.
b. An E fact table is created when activating an InfoCube with a number of partitions
corresponding to the partition value range.
c. If a cube is not partitioned before populating with data, it is not possible to partition the
cube without removing all the data.
d. PSA table partitions can contain more than one request.
e. All of the above.
ANSWER: E
The 'F' Fact table is partitioned on request and is created during data loads. The 'E' Fact
table of the InfoCube is created with the number of partitions specified in the partitioning
range. Once the data is loaded to the cube, partitioning based in the Time characteristic
is not possible without removing the data. PSA table partitions can contain more than
one request.