博客园_SAP ABAP/BW/BI

DSO参数调整的t-code

2012-05-15T03:47:00Z

RSODSO_SETTINGS

提高cube性能的一些方法（一）

2012-05-15T02:42:00Z

1. 尽量不要在Cube里放太detail的数据，这种需求首先考虑R3用ABAP解决，如果非要在BW，可以考虑在DSO出明细报表，在Cube出汇总报表，通过RRI接口调用明细报表。关于RRI，请看：
http://help.sap.com/saphelp_sm32/helpdata/en/99/08629bd3e41d418530c6849df303c9/content.htm

2. 当Cube的数据量很大时，可以拆分成多个Cube, 再用MultiProvider拼起来，这样query会在N个Cube中并行，提高效率。这就是所谓的逻辑分区。常见的分区方式有按年月，按国家，按BU，按类型等。

3. 对于很大的Cube，可以做partition, 这是物理分区，只支持按时间分区。

4. 使用Aggregation可以提高性能。但是Aggregation本身是cube的一个子集，提高性能的同时也加大了数据冗余，所以不要用太多。

5. 使用BIA是比Aggregation更有效的方法，就是要花不少钱。

6. 维度设计上，避免很多数据量很大char.放在一个维度上，因为这样会让维度表变得很大。通常，尽可能拆分成更多的维度，然后在 multiprovider层面，把相关的char都放一个维度里，然后做好Mapping，这样可以让用户更容易理解MultiProvider. 不过维度太多会导致fact table巨大，所以要做好平衡。

7. 对于material等很大的主数据，使用Line item Dimension. 此类Dimension只可以有一个Char.

8. 定期刷新DB Statistics 可以提高reporting的效率。

9. 给Cube做Compression。 Compression 本质上是去掉Data Dimension，这样fact table就被压缩了，但是request id 也消失了，将无法通过request id去管理数据。

本文链接

Process Chain\DTP\Infopackage相关的几个table和function module

2012-05-15T02:05:00Z

1、统计生产系统的PC

Table：
RSPCCHAINT
RSPCCHAINATTR

PC 日志表：
RSPCPROCESSLOG

FM：
RSPC_CHAIN_DELETE_REMOTE
unfreeze的FM：
RSPC_PROCESS_FINISH

2、统计P系统中PC包含的Variant
Table：
RSPCVARIANTATTR
RSPCVARIANT
RSPCVARIANTT
RSPCTRIGGER

FM：
RSPC_TRIGGER_DELETE
RSPC_VARIANT_DELETE

3、Infopackage

Table：
RSLDPIO

FM：
RSSM_SDL_INFOPACKAGE_DELETE

4、DTP

Table：
RSBKDTP
DTP_ORIGINAL

Report：
RSBKDTPDELETE

本文链接

Note 741478 - FAQ: Materialized views

2012-01-18T16:50:00Z

SAP Link	741478
Version	7
Release Status	Released for Customer
Released on	2010-11-30 12:05:35
Priority	Recommendations/additional info
Category	FAQ
Primary Component	BC-DB-ORA
Secondary Components	BC-DB-ORA-DBA

Summary
Symptom 1. What are materialized views? 2. What are the differences between views and materialized views? 3. What can I use materialized views for? 4. Are materialized views used in the SAP environment? 5. What are the prerequisites for materialized views? 6. How do I create a materialized view? 7. How can I refresh a materialized view? 8. How do I create a materialized view? 9. Which Oracle DDIC views exist for materialized views? 10. What internal objects are created that relate to materialized views? 11. How are materialized views used in online reorganizations? 12. What problems can occur with materialized views? 13. Where do I find more information about materialized views? Other terms This note provides answers to Frequently Asked Questions (FAQs). Solution 1. What are materialized views? Materialized views or snapshots are segments that contain the current or previous data of a table, a view or a join. You can update the data that was changed in the source objects since the creation of the materialized view immediately, at specific times or not at all. By default, materialized views as read-only. 2. What are the differences between views and materialized views? A view always returns the current information from the source objects while a materialized view can also contain a previous state of the source objects, depending on the refresh mode. In a regular view, the data must be retrieved from the source objects at runtime, which can lead to long runtimes for complex joins. A materialized view, however, can be compared with a table from which the system can read the data directly without overhead for joins. 3. What can I use materialized views for? With materialized views, you can replicate data automatically to other systems. The online table reorganization with DBMS_REDEFINITION is based on materialized views. You can cache expensive queries as materialized views and refresh them periodically (which can be compared with the table buffering in R/3). 4. Are materialized views used in the SAP environment? Materialized views are not used in the SAP environment. The only exception is the online table reorganization with BRSPACE which uses materialized views implicitly. Since materialized views are not used in the R/3 system, they cannot be created in the ABAP Dictionary. 5. What are the prerequisites for materialized views? By default, materialized views are based on primary key constraints. This means that a primary key constraint must be created on the table for this type of materialized view - otherwise ORA-12014 occurs. Alternatively, you can generate a materialized view based on the ROWID. If you want to refresh the materialized view automatically, you must set the JOB_QUEUE_PROCESSES parameter to 1 or a higher value. To create a materialized view, the database user must have the CREATE MATERIALIZED VIEW authorization. 6. How do I create a materialized view? You can create a simple materialized view with the following command: CREATE MATERIALIZED VIEW MY_MAT_VIEW <refresh_mode> AS <select_statement>; <select_statement> contains the query that is used to create the materialized view. Refer to the Oracle online documentation for more complex definitions of materialized views. The possible values for <refresh_mode> are listed in the answer to the next question. 7. How can I refresh a materialized view? There are multiple refresh options that you can specify in <refresh_mode>. First of all, there are FAST, COMPLETE and FORCE refresh operations: FAST: The data that you change in the source objects is logged in a log file for the materialized view. When you refresh, the system only needs to transfer the data to the materialized view. This method is used in the online reorganization. For a FAST refresh, the materialized view query must not be too complex and a log file for the materialized view must exist. COMPLETE: The materialized view is rebuilt completely during the refresh operation. FORCE: If possible, the system performs a FAST refresh, otherwise, it executes a COMPLETE refresh (default). In addition, you can specify a time for the refresh operation. For example: NEVER REFRESH: The refresh is not performed automatically. REFRESH FAST ON COMMIT: A FAST refresh is performed with every Commit. REFRESH COMPLETE START WITH SYSDATE NEXT SYSDATE + 1/48: A COMPLETE refresh is performed as of now every 30 minutes (30 minutes equals 1 day multiplied by 1/48). REFRESH FORCE START WITH TRUNC(SYSDATE) NEXT TRUNC(SYSDATE)+1/4: A FORCE refresh is performed at 12 am, 6 am, 12 pm and 6 pm. 8. How do I create a materialized view? You can use the following command to create a materialized view log for a table: CREATE MATERIALIZED VIEW LOG ON <table>; 9. Which Oracle DDIC views exist for materialized views? DBA_MVIEWS: Overview over the materialized views. DBA_MVIEW_LOGS: Overview over the log files for the materialized views. 10. What internal objects are created that relate to materialized views? If you create a primary key constraint that is not supported by an existing index as the basis for a materialized view, Oracle creates an index internally with the prefix SYS_C. When you create a materialized view log, the system creates the objects MLOG$_<table> and RUPD$_<table>. MLOG$_<table> is the materialized view log which records all changes to the underlying table. When you create a ROWID based materialized view, the system creates an I_SNAP$_<table> index for the materialized view, which receives the column M_ROW$$. 11. How are materialized views used in online reorganizations? If you use BRSPACE to execute an online reorganization based on the Oracle package DBMS_REDEFINITION (refer to Note 646681), then materialized views are also implicitly used due to the implementation of DBMS_REDEFINITION. In general, what happens is this: The target table with the name conention <table>#$ (see Note 541538) is treated as a materialized view of the source table <table>. The program creates a materialized view log for the source table <table>. This means that the system creates the objects MLOG$_<table> and RUPD$_<table> implicitly. If a primary index exists for the table to be reorganized, the online reorganization is executed on the basis of a primary key constraint. This means that the materialized view is also created on the basis of a primary key constraint. In turn, this means that an I_SNAP$_<table>#$ index is not necessary. If no primary index exists, the reorganization takes place on the basis of a ROWID materialized view, so the I_SNAP$_<table>#$ is implicitly created. 12. What problems can occur with materialized views? Problems with the table owner's default tablespace All implicit objects such as MLOG$_<table>, RUPD$_< table> and I_SNAP$_<table>#$ are created in the table owner's default tablespace. This may cause tablespace overflows and MAXEXTENTS problems under some circumstances, for example: ORA-12008: error in materialized view refresh path ORA-01632: max # extents (505) reached in index SAPR3.I_SNAP$_<table> ORA-01652: unable to extend temp segment in tablespace <tsp> Therefore, take not of the following during the online reorganization: Ensure that there is sufficient free space available in the table owner's default tablespace, and that problems relating to MAXEXTENTS can be avoided (such as by setting MAXEXTENTS -> UNLIMITED). Avoid reorganizing large tables that have no primary index, since this means that the program must create large I_SNAP$ indexes implicitly. Avoid online reorganizations in parallel with massive data chagnes to the tables to be reorganized, since otherwise the MLOG$ table grows very large. Increased Redo log count For each periodic COMPLETE refresh of a materialized view, the system creates Redo information which can lead to a significant Redo log count. ORA-12014: table does not contain a primary key constraint If you create a primary key materialized view for a table without primary key constraint, the system issues an ORA-12014 error message. You can create a primary key constraint with the following command: ALTER TABLE <table> ADD PRIMARY_KEY (<column_list>); ORA-12028: materialized view type is not supported by master site The system issues this error when you create materialized views with an Oracle databases Release 7 or earlier. If the system issues the error message when you reorganize a database version 9i or later online, parts of the Oracle DDICs may not have a current status yet or old Oracle packages may still exist under users such as SAPR3 or SAP<sid>. ORA-12091: cannot online redefine table with materialized views If a materialized view or a log file for a materialized view already exists for a table, an online reorganization will lead to an ORA-12091 error message. The error occurs if the online reorganization for a table is stopped improperly and you then try to restart the online reorganization. To solve the problem, you must clean up the rest of the previous online reorganization as follows: brspace -u / -f tbreorg -t <table_name> -a cleanup If you want to clean up the remainder of a large quantity of tables, you can also use "-t". However, this causes long runtime for the cleanup. ORA-12092: cannot online redefine replicated table You are trying to reorganize a materialized view online. This is not possible. ORA-12096: error in materialized view log on <table> If problems occur when you try to access a log file for a materialized view, the system issues an ORA-12096 error message which is followed by the actual error. If the ORA-12096 error message is accompanied by an ORA-00942 message, the problem may be due to an online reorganization that was only partially undone. In this case, you can drop the log file for the materialized view manually (after you have verified that it is no longer required): DROP MATERIALIZED VIEW LOG ON <table>; ORA-23413: table does not have a materialized view log If you try to create a FAST materialized view for a table without materialized view log, the system issues an ORA-23413 error message. The name of the materialized view log is always MLOG$_<table> and you can create the log with the following command: CREATE MATERIALIZED VIEW LOG ON <table>; 13. Where do I find more information about materialized views? Refer to the Oracle online documentation for detailed information about materialized views: Oracle9i Advanced Replication -> 3 Materialized view Concepts and Architecture Oracle9i SQL Reference -> CREATE MATERIALIZED VIEW Refer to note 646681 for information about the online reorganization with BRSPACE

Summary

Symptom

1. What are materialized views?

2. What are the differences between views and materialized views?

3. What can I use materialized views for?

4. Are materialized views used in the SAP environment?

5. What are the prerequisites for materialized views?

6. How do I create a materialized view?

7. How can I refresh a materialized view?

8. How do I create a materialized view?

9. Which Oracle DDIC views exist for materialized views?

10. What internal objects are created that relate to materialized views?

11. How are materialized views used in online reorganizations?

12. What problems can occur with materialized views?

13. Where do I find more information about materialized views?

Other terms

This note provides answers to Frequently Asked Questions (FAQs).

Solution

1. What are materialized views?

Materialized views or snapshots are segments that contain the current or previous data of a table, a view or a join. You can update the data that was changed in the source objects since the creation of the materialized view immediately, at specific times or not at all.

By default, materialized views as read-only.

2. What are the differences between views and materialized views?

A view always returns the current information from the source objects while a materialized view can also contain a previous state of the source objects, depending on the refresh mode.

In a regular view, the data must be retrieved from the source objects at runtime, which can lead to long runtimes for complex joins. A materialized view, however, can be compared with a table from which the system can read the data directly without overhead for joins.

3. What can I use materialized views for?

With materialized views, you can replicate data automatically to other systems.

The online table reorganization with DBMS_REDEFINITION is based on materialized views.

You can cache expensive queries as materialized views and refresh them periodically (which can be compared with the table buffering in R/3).

4. Are materialized views used in the SAP environment?

Materialized views are not used in the SAP environment. The only exception is the online table reorganization with BRSPACE which uses materialized views implicitly.

Since materialized views are not used in the R/3 system, they cannot be created in the ABAP Dictionary.

5. What are the prerequisites for materialized views?

By default, materialized views are based on primary key constraints. This means that a primary key constraint must be created on the table for this type of materialized view - otherwise ORA-12014 occurs. Alternatively, you can generate a materialized view based on the ROWID.

If you want to refresh the materialized view automatically, you must set the JOB_QUEUE_PROCESSES parameter to 1 or a higher value.

To create a materialized view, the database user must have the CREATE MATERIALIZED VIEW authorization.

6. How do I create a materialized view?

You can create a simple materialized view with the following command:

CREATE MATERIALIZED VIEW MY_MAT_VIEW
<refresh_mode>
AS <select_statement>;

<select_statement> contains the query that is used to create the materialized view.

Refer to the Oracle online documentation for more complex definitions of materialized views. The possible values for <refresh_mode> are listed in the answer to the next question.

7. How can I refresh a materialized view?

There are multiple refresh options that you can specify in <refresh_mode>. First of all, there are FAST, COMPLETE and FORCE refresh operations:

FAST: The data that you change in the source objects is logged in a log file for the materialized view. When you refresh, the system only needs to transfer the data to the materialized view. This method is used in the online reorganization. For a FAST refresh, the materialized view query must not be too complex and a log file for the materialized view must exist.

COMPLETE: The materialized view is rebuilt completely during the refresh operation.

FORCE: If possible, the system performs a FAST refresh, otherwise, it executes a COMPLETE refresh (default).

In addition, you can specify a time for the refresh operation. For example:

NEVER REFRESH: The refresh is not performed automatically.

REFRESH FAST ON COMMIT: A FAST refresh is performed with every Commit.

REFRESH COMPLETE START WITH SYSDATE NEXT SYSDATE + 1/48: A COMPLETE refresh is performed as of now every 30 minutes (30 minutes equals 1 day multiplied by 1/48).

REFRESH FORCE START WITH TRUNC(SYSDATE) NEXT TRUNC(SYSDATE)+1/4: A FORCE refresh is performed at 12 am, 6 am, 12 pm and 6 pm.

8. How do I create a materialized view?

You can use the following command to create a materialized view log for a table:

CREATE MATERIALIZED VIEW LOG ON <table>;

9. Which Oracle DDIC views exist for materialized views?

DBA_MVIEWS: Overview over the materialized views.

DBA_MVIEW_LOGS: Overview over the log files for the materialized views.

10. What internal objects are created that relate to materialized views?

If you create a primary key constraint that is not supported by an existing index as the basis for a materialized view, Oracle creates an index internally with the prefix SYS_C.

When you create a materialized view log, the system creates the objects MLOG$_<table> and RUPD$_<table>. MLOG$_<table> is the materialized view log which records all changes to the underlying table.

When you create a ROWID based materialized view, the system creates an I_SNAP$_<table> index for the materialized view, which receives the column M_ROW$$.

11. How are materialized views used in online reorganizations?

If you use BRSPACE to execute an online reorganization based on the Oracle package DBMS_REDEFINITION (refer to Note 646681), then materialized views are also implicitly used due to the implementation of DBMS_REDEFINITION. In general, what happens is this:

The target table with the name conention <table>#$ (see Note 541538) is treated as a materialized view of the source table <table>.

The program creates a materialized view log for the source table <table>. This means that the system creates the objects MLOG$_<table> and RUPD$_<table> implicitly.

If a primary index exists for the table to be reorganized, the online reorganization is executed on the basis of a primary key constraint. This means that the materialized view is also created on the basis of a primary key constraint. In turn, this means that an I_SNAP$_<table>#$ index is not necessary.

If no primary index exists, the reorganization takes place on the basis of a ROWID materialized view, so the I_SNAP$_<table>#$ is implicitly created.

12. What problems can occur with materialized views?

Problems with the table owner's default tablespace

All implicit objects such as MLOG$_<table>, RUPD$_< table> and I_SNAP$_<table>#$ are created in the table owner's default tablespace. This may cause tablespace overflows and MAXEXTENTS problems under some circumstances, for example:

ORA-12008: error in materialized view refresh path
ORA-01632: max # extents (505) reached in index
SAPR3.I_SNAP$_<table>
ORA-01652: unable to extend temp segment in tablespace <tsp>

Therefore, take not of the following during the online reorganization:

Ensure that there is sufficient free space available in the table owner's default tablespace, and that problems relating to MAXEXTENTS can be avoided (such as by setting MAXEXTENTS -> UNLIMITED).

Avoid reorganizing large tables that have no primary index, since this means that the program must create large I_SNAP$ indexes implicitly.

Avoid online reorganizations in parallel with massive data chagnes to the tables to be reorganized, since otherwise the MLOG$ table grows very large.

Increased Redo log count

For each periodic COMPLETE refresh of a materialized view, the system creates Redo information which can lead to a significant Redo log count.

ORA-12014: table does not contain a primary key constraint

If you create a primary key materialized view for a table without primary key constraint, the system issues an ORA-12014 error message.

You can create a primary key constraint with the following command:

ALTER TABLE <table> ADD PRIMARY_KEY (<column_list>);

ORA-12028: materialized view type is not supported by master site

The system issues this error when you create materialized views with an Oracle databases Release 7 or earlier. If the system issues the error message when you reorganize a database version 9i or later online, parts of the Oracle DDICs may not have a current status yet or old Oracle packages may still exist under users such as SAPR3 or SAP<sid>.

ORA-12091: cannot online redefine table with materialized views

If a materialized view or a log file for a materialized view already exists for a table, an online reorganization will lead to an ORA-12091 error message. The error occurs if the online reorganization for a table is stopped improperly and you then try to restart the online reorganization. To solve the problem, you must clean up the rest of the previous online reorganization as follows:

brspace -u / -f tbreorg -t <table_name> -a cleanup

If you want to clean up the remainder of a large quantity of tables, you can also use "-t". However, this causes long runtime for the cleanup.

ORA-12092: cannot online redefine replicated table

You are trying to reorganize a materialized view online. This is not possible.

ORA-12096: error in materialized view log on <table>

If problems occur when you try to access a log file for a materialized view, the system issues an ORA-12096 error message which is followed by the actual error. If the ORA-12096 error message is accompanied by an ORA-00942 message, the problem may be due to an online reorganization that was only partially undone. In this case, you can drop the log file for the materialized view manually (after you have verified that it is no longer required):

DROP MATERIALIZED VIEW LOG ON <table>;

ORA-23413: table does not have a materialized view log

If you try to create a FAST materialized view for a table without materialized view log, the system issues an ORA-23413 error message. The name of the materialized view log is always MLOG$_<table> and you can create the log with the following command:

CREATE MATERIALIZED VIEW LOG ON <table>;

13. Where do I find more information about materialized views?

Refer to the Oracle online documentation for detailed information about materialized views:

Oracle9i Advanced Replication
-> 3 Materialized view Concepts and Architecture

Oracle9i SQL Reference
-> CREATE MATERIALIZED VIEW

Refer to note 646681 for information about the online reorganization with BRSPACE

本文链接

SAP BW Query常见的性能问题

2011-12-16T07:26:00Z

BW Query Performance

Question:
1. What kind of tools are available to monitor the overall Query Performance?

Answers:
o BW Statistics
o BW Workload Analysis in ST03N (Use Export Mode!)
o Content of Table RSDDSTAT

Question:
2. Do I have to do something to enable such tools?

Answer:
o Yes, you need to turn on the BW Statistics:
RSA1, choose Tools -> BW statistics for InfoCubes
(Choose OLAP and WHM for your relevant Cubes)

Question:
3. What kind of tools are available to analyse a specific query in detail?

Answers:
o Transaction RSRT
o Transaction RSRTRACE

Question:
4. Do I have a overall query performance problem?

Answers:
o Use ST03N -> BW System load values to recognize the problem. Use the
number given in table 'Reporting - InfoCubes:Share of total time (s)'
to check if one of the columns %OLAP, %DB, %Frontend shows a high
number in all InfoCubes.
o You need to run ST03N in expert mode to get these values

Question:
5. What can I do if the database proportion is high for all queries?

Answers:
Check:
o If the database statistic strategy is set up properly for your DB platform
(above all for the BW specific tables)
o If database parameter set up accords with SAP Notes and SAP Services (EarlyWatch)
o If Buffers, I/O, CPU, memory on the database server are exhausted?
o If Cube compression is used regularly
o If Database partitioning is used (not available on all DB platforms)

Question:
6. What can I do if the OLAP proportion is high for all queries?

Answers:
Check:
o If the CPUs on the application server are exhausted
o If the SAP R/3 memory set up is done properly (use TX ST02 to find
bottlenecks)
o If the read mode of the queries is unfavourable (RSRREPDIR, RSDDSTAT,
Customizing default)

Question:
7. What can I do if the client proportion is high for all queries?

Answer:
o Check whether most of your clients are connected via a WAN Connection and the amount
of data which is transferred is rather high.

Question:
8. Where can I get specific runtime information for one query?

Answers:
o Again you can use ST03N -> BW System Load
o Depending on the time frame you select, you get historical data or
current data.
o To get to a specific query you need to drill down using the InfoCube
name
o Use Aggregation Query to get more runtime information about a
single query. Use tab All data to get to the details.
(DB, OLAP, and Frontend time, plus Select/ Transferred records,
plus number of cells and formats)

Question:
9. What kind of query performance problems can I recognize using ST03N
values for a specific query?

Answers:
(Use Details to get the runtime segments)
o High Database Runtime
o High OLAP Runtime
o High Frontend Runtime

Question:
10. What can I do if a query has a high database runtime?

Answers:
o Check if an aggregate is suitable (use All data to get values
"selected records to transferred records", a high number here would
be an indicator for query performance improvement using an aggregate)
o Check if database statistics are update to data for the
Cube/Aggregate, use TX RSRV output (use database check for statistics
and indexes)
o Check if the read mode of the query is unfavourable - Recommended (H)

Question:
11. What can I do if a query has a high OLAP runtime?

Answers:
o Check if a high number of Cells transferred to the OLAP (use
"All data" to get value "No. of Cells")
o Use RSRT technical Information to check if any extra OLAP-processing
is necessary (Stock Query, Exception Aggregation, Calc. before
Aggregation, Virtual Char. Key Figures, Attributes in Calculated
Key Figs, Time-dependent Currency Translation)
together with a high number of records transferred.
o Check if a user exit Usage is involved in the OLAP runtime?
o Check if large hierarchies are used and the entry hierarchy level is
as deep as possible. This limits the levels of the
hierarchy that must be processed. Use SE16 on the inclusion
tables and use the List of Value feature on the column successor
and predecessor to see which entry level of the hierarchy is used.
- Check if a proper index on the inclusion table exist

Question:
12. What can I do if a query has a high frontend runtime?

Answers:
o Check if a very high number of cells and formattings are transferred
to the Frontend ( use "All data" to get value "No. of Cells") which
cause high network and frontend (processing) runtime.
o Check if frontend PC are within the recommendation (RAM, CPU Mhz)
o Check if the bandwidth for WAN connection is sufficient

本文链接

SAP BW维护项目中的常见问题

2011-12-16T07:15:00Z

Below are some of the errors in a support project which will be a great help for new learners:

1) RFC connection lost.
A) We can check out in the SM59 t-code
RFC Destination
+ R/3 connection
CRD client (our r/3 client)
double click..test connection in menu

2) Invalid characters while loading.
A) Change them in the PSA & load them.

3) ALEREMOTE user is locked.
1) Ask your Basis team to release the user. It is mostly ALEREMOTE.
2) Password Changed
3) Number of incorrect attempts to login into ALEREMOTE.
4) USE SM12 t-code to find out are there any locks.

4) Lower case letters not allowed.
A) Uncheck the lower case letters check box under "general" tab in the info object.

5) While loading the data I am getting messeage that 'Record the field mentioned in the errror message is not mapped to any infoboject in the transfer rule'.
A)

6) Object locked.
A) It might be locked by some other process or a user. Also check for authorizations

7) "Non-updated Idocs found in Source System".
A) Check whether any TRFC s strucked in source system. You can check it in SM58. If no TRFC s are there then better trigger the load again i.e., change the status to red, delete the bad request and run the load. Check whether the load is Delta or Full. If it is full just go ahead with the above step.

If it is Delta check wheteher it is source system or BW. If it is source system go for repeat delta. If it is BW then you need to reset Data Mart status.

8) While loading master data, one of the datapackage has a red light error message: Master data/text of characteristic ZCUSTSAL already deleted.
A)

9) Extraction job aborted in r3
A) It might have got cancelled due to running for more than the expected time, or may be cancelled by R/3 users if it is hampering the performance.

10) Request couldn't be activated because there is another request in the psa with a smaller sid
A)

11) Repeat of last delta not possible.
A) Repeat of last delta is not a option, but a mandate, in case the delta run failed. In such a case, we cant run the simple delta again. The system is going to run a repeat of last delta, so as to collect the failed delta's data again as well as any data collect till now right from failure.

For repeat of last delta to be run, we should have the previous delta failed. Lets assume, in your case, I am not sure, if the delta got falied or deleted. If this is a deletion, then we need to catch hold of the request and make the status to red. This is going to tell the system that the delta failed(although it ran successfully, but you are forcing this message to the system). Now, if you run the delta info package, it will fetch the data related to 22nd plus all the changes from there on till today.

An essential point here, you should not have run any deltas after 22nd till now. Then only repeat of last delta is going to work. Otherwise only option is to run a repair full request with data selections, if we know selection parameters.

12) Datasource not replicated
A) Replicate the datasource from R/3 through source system in the AWB & assign it to the infosource and activate it again.

13) Datasource/transfer structure not active.
A) Use the function module RS_TRANSTRU_ACTIVATE_ALL to activate it

14) ODS activation error.
A) ODS activation errors can occur mainly due to following reasons-
1.Invalid characters (# like characters)
2.Invalid data values for units/currencies etc
3.Invalid values for data types of char & key figures.
4.Error in generating SID values for some data.

本文链接

SAP BW实施项目包含哪些内容,需要跟R/3系统进行交互

2011-12-16T07:05:00Z

What type of activities we perform in IMPLEMENTATION project? Will we start from modeling stage or interact with R/3?

Responsibilities of an implementation project

For ex, Lets say If its a fresh implementation of BI or for that matter you are implementing SAP.

First and foremost will be your requirements gathering from the client. Depending upon the requirements you will creat a business blueprint of the project which is the entire process from the start to the end of an implementation.

After the blue print phase sign off we start off with the realization phase where the actual development happens. In our example after installing the necessary softwares, patches for BI we need to discuss with the end users who are going to use the system for inputs like how they want a report to look like and what are the Key Performance Indicators(KPI) for the reports etc., basically its a question and answer session with the business users.

After collecting those informations the development happens in the development servers.

After the development comes to an end the same objects are tested in quality servers for any bugs, errors etc., When all the tests are done we move all the objects to the production environment and test it again whether everything works fine.

The Go-Live of the project happens where the actually postings happen from the users and reports are generated based on those inputs which will be available as an analytical report for the management to take decisions.

The responsibilites vary depending on the requirement. Initially the business analyst will interact with the end users/managers etc., then on the requirements the software consultants do the development, testers do the testing and finally the go-live happens.

BW Data Architect Description:
The BW Data Architect is responsible for the overall data design of the BW project. This includes the design of the:
BW InfoCubes (Basic Cubes, Multi-cubes, Remote cubes, and Aggregates) » BW ODS Objects » BW Datamarts » Logical Models » BW Process Models » BW Enterprise Models

The BW Data Architect plays a critical role in the BW project and is the link between the end user's business requirements and the data architecture solution that will satisfy these requirements. All other activities in the BW project are contingent upon the data design being sound and flexible enough to satisfy evolving business requirements.

Time Commitment
– the time which must be committed to this Role to ensure the project requirements are met Project Complexity Time Commitment Low If the BW project utilizes standard BW content and InfoCubes, this role can be satisfied by the BW Application Consultant.

Medium
If the BW project requires enhancements to the standard BW content and InfoCubes and/or requires the integration of non-SAP data, this role may require a committed resource.

High
If the BW project requires significant modification and enhancement to standard BW content and InfoCubes, it is highly recommended that an experienced resource be committed full-time to the project.

Key Attributes
The BW Data Architect must have:
An understanding of the BW data architecture » An understanding of multidimensional modeling » An understanding of the differences between operational systems data modeling and data warehouse data modeling » An understanding of the end user's data » An understanding of the integration points of the data (e.g., customer number, invoice number) » Excellent troubleshooting and analytical skills » Excellent communication skills » Technical competency in data modeling » Multi-language skills, if an international implementation » Working knowledge of the BW and R/3 application(s) » Experience with Data Modeling application software (i.e., ERWIN, Oracle Designer, S-Designer, etc.)

Key Tasks
The BW Data Architect is responsible for capturing the business requirements for the BW project. This effort includes:
Planning the business requirements gathering sessions and process » Coordinating all business requirements gathering efforts with the BW Project Manager » Facilitating the business requirements gathering sessions » Capturing the information and producing the deliverables from the business requirements gathering sessions » Understanding and documenting business definitions of data » Developing the data model » Ensuring integration of data from both SAP and non-SAP sources » Fielding questions concerning the data content, definition and structure

This role should also address other critical data design issues such as:
Granularity of data and the potential for multiple levels of granularity » Use of degenerate dimensions » InfoCube partitioning » Need for aggregation at multiple levels » Need for storing derived BW data » Ensuring overall integrity of all BW Models » Providing Data Administration development standards for business requirements analysis and BW enterprise modeling » Provide strategic planning for data management » Impact analysis of data change requirements As stated above, the BW Data Architect is responsible for the overall data design of the BW project.

This includes the design of the:
BW InfoCubes (Basic Cubes, Multi-cubes, Remote cubes, and Aggregates) » BW ODS Objects » BW Datamarts » Logical Models » BW Process Models » BW Enterprise Models

BW Project Implementation and Rollout

Global IDs uses a modified version of the ASAP methodology to ensure a robust BW environment is implemented, tested, and delivered. The project milestones include the following:

Blueprinting
Realization
Roll Out
Post-Implementation Support
Organizational Change Management
BW Maintenance

*Phase 1: Upgrade Project Preparation -*

The purpose of this phase is to provide initial planning and preparation for your R/3 upgrade project. Although each upgrade project has unique objectives, scope and priorities, the steps in Phase 1 help identify and plan the primary topics that must be considered.

*Phase 2: Business Blueprint -*

The purpose of this phase is to establish a common understanding in your company about how the new R/3 release will handle your current business processes after the R/3 upgrade.

*Phase 3: Realization of DEV and QAS -*

The purpose of this phase is to perform the upgrade of all non-production R/3 Systems, which normally include development (DEV) and quality assurance
(QAS) systems, and to ensure the quality of this upgrade process by comprehensive testing. You need to upgrade your DEV and QAS systems before performing the upgrade for your production system (PRD). Furthermore, you will develop your end user documentation and training in this phase.

*Phase 4: Realization of Production -*

The purpose of this phase is to complete the final preparation for the upgrade of the production R/3 System, and end user training, if necessary. On the successful completion of this phase, you are ready to run your business in your upgraded production R/3 System.

*Phase 5: Go Live and Support -*

The purpose of this phase is the transition from your previous R/3 System to a newly upgraded and improved R/3 System. Be aware that you must set up an ongoing support organization to service your R/3 users, both during the first few days of your upgraded production operations, and on an ongoing basis.

本文链接

SAP BW 系统管理员是一个什么样的角色,日常工作是什么,具体该掌握一些什么知识

2011-12-16T06:54:00Z

What is the role of consultant has to play if the title is BW administrator? What is his day to day activity and which will be the main focus area for him in which he should be proficient?

BW Administartor - is the person who provides Authorization access to different Roles, Profiles depending upon the requirement.

For eg. There are two groups of people : Group A and Group B.

Group A - Manager

Group B - Developer

Now the Authorization or Access Rights for both the Groups are different.

So for doing this sort of activity.........we required Administrator.

本文链接

BW Technical and Functional 的区别是什么

2011-12-16T06:52:00Z

Difference Between BW Technical and Functional

In general Functional means, derive the funtional specification from the business requirement document. This job normally is done either by the business analyst or system analyst who has a very good knowledge of the business. In some large organizations there will be a business analyst as well as system analyst.

In any business requirement or need for new reports or queries originates with the business user. This requirement will be recorded after discussion by the business analyst. A system analyst analyses these requirements and generates functional specification document. In the case of BW it could be also called logical design in DATA MODELING.

After review this logical desing will be translated to physical design . This process defines all the required dimensions, key figures, master data, etc.

Once this process is approved and signed off by the requester(users), then conversion of this into practically usable tasks using the SAP BW software. This is called Technical. The whole process of creating an InfoProvider, InfoObjects, InforSources, Source system, etc falls under the Technical domain.

本文链接

SAP BW中采购的业务报表有哪些

2011-12-16T06:46:00Z

Give details of any standard BI report that are relevant to Purchasing?

Answer:

For any Reports you just go to RSRREPDIR table and execute it and see the cube wise/module wise reports.

Eg: Give cube = 0PUR* then execute it will display all reports. Then take COMPUID goto RSZELTTXT table and give this ID and get description.

These are some of the BI standard reports for purchasing:

Contract Details : Technical Name: 0SRCT_DS1_Q003

Consolidated Purchase Order Value Analysis (Over Three Months) : Technical Name: 0BBP_C01_Q029

ABC Analysis :Technical name: 0BBP_C01_Q041

Procurement Values with/Without Contracts : Technical Name: 0BBP_C01_Q020

Expiring Contracts : Technical Name: 0SRCT_DS1_Q004

Quantity reliability :Technical Name: 0BBP_C01_Q011

Procurement Values per Service Provider :Technical name: 0BBP_C01_Q019

Delivery Delay of Last Confirmation :Technical Name: 0BBP_C01_Q012

Procurement Card Use :Technical Name: 0BBP_C01_Q015

Procurement Values per Vendor : Technical Name: 0BBP_C01_Q005

Procurement Values per Product Category :Technical Name: 0BBP_C01_Q004

本文链接