Design of process chain

Hi guys
I have a cube where i am loading data from 3 DSO's in which 2 DSO's are using a standarad data source and is used by many cubes and the 3rd dso is a generic datasource which i created for my requirement..and the standarad dso's info pack have been already put in the process chain..now how should i design my process chain in order to get the data loaded in to my cube..

Hi,
create InfoPackage (I think you are loading PSA and data target in parallel) to load data into our 3rd DSO. Create process for activating the data in the DSOs..
Then load in the Cube (by DTP), build Indices (for the Cube if you have droped them), build DB statistics and roll up the aggregates for your InfoCube.
Check:
[Using Process Chains in SAP Business Information Warehouse|https://www.sdn.sap.com/irj/scn/go/portal/prtroot/docs/library/uuid/b0afcd90-0201-0010-b297-9184845346ca]
Regards
Andreas

Similar Messages

Error while designing the Process Chains

Hi All,
I have a issue on Process Chains.
In my Project there are 2 ODS.
ODS 1 : Status Header ODS (ZODS)
ODS 2 : Service Order Header. (0CRM_PROH)
In CRM, if any Service Order is created, it is available in delta Queue(RSA7) immediately.
While loading the data to ODS 2(Service Order),it is checking the data from status ODS 1(Status ODS)
For Example, I am loading the data of service Order 80001 to ODS 2(0CRM_PROH). Before loading to ODS 2, i am checking whether this service Order Number(80001) is available in ODS 1.
If the service Order 80001 is there in ODS 1, it will update the data to ODS 2.
If the service order 80001 is not there in ODS 1, it will not Update the data to ODS 2.
Problem i am facing :
1. First i am loading the data to ODS 1 (Status ODS). Early Morning 2.30AM and it completes the data load at 2.40AM. And document posting is happening on 2.37AM and ODS 1 doesnt pull the data of document posted.
2. Once the ODS 1 is completed, by 2.41AM, the data loading will happen for ODS 2 (Service Order Header).It brings all the data including the document posted on 2.37AM.
3.While loading all the data is available, and while loading to ODS 2, it is checking the status ODS. And it is checking for the document posted on 2.37AM and this document is not there in Status ODS 1 and the data(posted on 2.37AM) will not update in ODS 2.
Because of this, some records are missing in ODS 2. And while loading, we cant stop the document posting also.
How can i achieve this using the process Chain.
Design i have tried in Process Chain :
1. Start Variant.
2. Load Infopackage to PSA only for ODS 2.
3. Load Infopackage to PSA and activate ODS for ODS 1 (Status ODS)
4. Read PSA (of ODS 2) to Load data to ODS 2.
5. Activate the ODS.
I tried, but i got a error message while designing the process Chains.
No type "Activate DataStore Object Data" process can be in front of process "Read PSA and Update Data Target" variable PSA_SERVICE_ORDER in the chain
Pls suggest how to recify this according to my design.
Thanks & Regards,
Ram.

Hi,
You can try this..
.......................................... Start ...................................................
Load data to PSA (ODS 2) ..................................... Load data to PSA (ODS 1)
..................................................................................Load data to ODS 1
...................................................................................Activate ODS 1
........................................... AND ..........................................................
........................................ Load data to ODS 2 .............................................
.......................................... Activate ODS 2 ...........................................
Regards,
Anup.
Edited by: Anup ShankarBS on Sep 3, 2009 4:56 PM

On what Basis V need to Design the Process Chain in BI7

Dear All,
Hope every one having a great time. I am designing the Process Chain for the DTP's in BI7 its for the Attributes load, what things i need to consider while designing dis chain mean is der any dependency i need to follow suppose der is "material" is an info provider and it consist of "material type" as attribute, where dis "material type' is also a infoprovider which one should i load first. and how many processes can i assign to start variant n what is the effective way to decide dis. shall i need to consider the Datasource also suppose if the 2 or 3 process r feeding the data from the same Datasource den what i hav to do in BI7.
Your Contribution will b appreciated
Thanx in Advance.
KK

Hi ,
There are several schools of thought on how to design your process chain for master data. For us, we load the attributes first, then hierarchies, then run ACR, and then load the texts.
We have a meta chain (same as a process chain) that executes 2 process chains. The first loads our attibutes, hierarchies and runs the ACR. The second loads the texts. So, theoretically, we have one process chain that loads all our master data.
Things to consider:
1. Determine how many batch processes you can execute at one time on the system. This will determine how many infopackages you can have running at the same time (big influence on the design of your process chain).
2. Determine the order in which you want to update your master data. Some say to update the infoobjects that are attributes of other master data first. Also, some extracts may take longer to execute than others. This may also influence your design.
3. When connecting the infopackages in the process chain, you will be prompted as to how the connection should proceed (successful, error, or always). In your case, the decision should be always.
Pls search for the links like this;
http://help.sap.com/saphelp_nw04/helpdata/en/ce/dc87c1d711f846b34e0e42ede5ebb7/content.htm
Re: Next step...
Hope this helps.
Regards
CSM Reddy

Pharma Detailed Design Documents - Process Chains, BEx, Extractors, Queries

Hello,
I was wondering if I can get a template for the Design Documents for Process Chains , Business Explorer (BEx) , Queries , WAD Reports , and Extractors that cover the features in NetWeaver 2004s, from anyone working in the Pharmaceutical industry. All help will be greatly appreciated. Please send emails to [email protected]
Regards
Sam

Hi!
I think main difference in pharma industry will be in Business partner having terriotires and bricks. just check the following link
http://help.sap.com/saphelp_nw04s/helpdata/en/cb/e11540bf1af72ee10000000a1550b0/frameset.htm
regards
mohan

Error while designing the process Chain.

Hi All,
I have created the Process Chain.
1. Start Variant.
2. Load data to PSA for ODS 1.
3. Load data to ODS 2
4. Activate the ODS 2
5. Read PSA and load data target to ODS 1.
6. Activate the ODS 1.
While checking the process Chain, i am getting the error message as
"No Type "Activate Datastore Object data" process can be in front of process "Read PSA and data target" Variable PSA_CONFIRM_ITEM in the chain.
Can u pls help on this issue.
Thanks,
Jelina.

Hi,
You can try this..
..................................................... Start ...................................................
Load data to PSA for ODS 1 ..................................Load data to ODS 2
...................................................................................Activate ODS 2
.............................................. AND ..........................................................
........................ Read PSA and load data To ODS 1 .............................................
.......................................... Activate ODS 1 ............................................
Regards,
Anup.
Edited by: Anup ShankarBS on Sep 4, 2009 10:47 AM

Parallel Process in a Process chain design

Hi
BAsed on what factors can we make a decision on How many parallel data loads (process) we can include while designing a Process chains
Thanks

Hi Maxi,
There is no hard and fast rule for that, for trial purpose you can add specific no. of parallel processes and schedule the chain, if there are not enough background processes available to fulfill your request then SAP will give you warning there you can see how many processes are available.
But if you go for maximum parallel process then it actually depends on how many processes are available at the time of process chain scheduling. Though your server have enough process but if they are utilized by other processes then still you will get warning while executing process chain.
So just check in your server how many background processes are there and then take some optimum decision.
Regards,
Durgesh.

Process Chains Design

Hi Experts
I need to design a process chain .The data flow is
Flat File --> ODS ---> CUBE
Could you please let me know the detailed steps involved
Like...start variant,info package (ods) activation ods requ, drop cube index, info package (cube...etc
I need the detailes list of processes to be used for that flow and why we use them as i need to explain same to my manager.
I shall be greatfull if you can send any screen shots to my mail id...[email protected]
Thanks in advance

Hi,
Here are the steps :
1. Start variant to schdule the process chain
2. Goto Process Type Load Process and Post Preocessing. Drag execute infopackage with it. If it comes with execute attribute change run then just get rid of it. This one is for loading your ODS from flat file. Other way to do it goto Infosources tab and then select the one that is the flat file load. Expand it and drag the flat file infopackage to the right hand window. Link the start variant to the the infopackage load.
3. Now you need to activate the ODS request. Goto Process Type and goto Data Target administration. Select Activate ODS Object data and drag it to the right hand window. Link the ODS load to this process type. Define condition when linking that this step will execute only if the ODS load is successful.
4. Then you need to select the Process type Further Processing to update the ODS data further.
5. When you load the delta to the Cube after this, you need to decide whether to use the process types delete indexes and then recreating them after the load. If the delta is very small then no need for these steps, if delta is large, then next step would be to drop indexes ->load cube->recreate indexes. This helps data load performance. If you have aggregates on the cube, next step would be to select the Process type for Roll up of Aggregates.
Cheers,
Kedar

Questions on process chains

Hai all,
I have two questions on Process chains.
1. I created a process chain for master data and a process chain for data loads(with ODS and cube). As the Master data has to be loaded first before loading transaction data, can I include the Master data process chain as a local process chain soon after the start process and then only after loading master data, the system proceeds to transaction data?
2. I designed a process chain with the aggreagtion and compression of cube data. I have forgot to check the option delete overlapping requests in the infopackage to load the infocube. I ran the process chain and now there are 2 requests and the duplicate documents. I want to delete the recent request and check the deleting duplicate requests option in info package and re run. The problem is the system is not allowing me to delete the recent request.
I appreciate any kind of help.
Thanks.

Hai Bhanu,
Thanks for the reply. I am sceduling the request for deletion in the Manage infocube> but the request is not getting deleted. Everytime I click refresh, the delete icon disappears and the screen becomes as it was before scheduing the deletion. I checked after sometime, even then the same thing...
I wonder the collpased requests can not be deleted???

Process chain questions

Hai All,
           I designed a process chain with data flow like this:
R/3>ODS1>ODS2>Cube.
The process chain sequence is:
Start>>Load data>>activate ODS1 data>>load data to ODS2>>activate ODS2 data>>delete index>>load data to cube.
           My process chain is working and the data loads are good. But I am getting warning messages for all three variants "activate ODS1 data>>load data to ODS2>>activate ODS2 data>>"
The message for the two activation variants is as follows:
A type "Update ODS Object Data (Further Update)" process has to follow process "Activate ODS Object Data" var.ACTIVATEZMONDET in the chain
Message no. RSMPC016
The message for the loading variant is as follows:
A type "Activate ODS Object Data" process cannot precede process "Execute InfoPackage" var. ZPAK_3YF4Q8R0RSHTDJH74WZOAXBC6 in thechain
Message no. RSMPC013
           The version I am working on is BW 3.5. My question is it ok to have these warning messages or do I need to do anything else?
Thank you.

Hi Visu,
If you data loads are running fine, you can ignore the warning messages. The system tries to warn you in case it suspects that a process is missing from the chain of events required...but there are more than one ways to achieve a load. If the data volume to be loaded to the cube is not very high, like it is just a daily delta, you can even go ahead and remove the index deletion process.
Hope this helps...

Creating cube index in process chain

Hi,
From my previous post I realized that we build index of cube first and then delete the overlapping request from cube.
(http://help.sap.com/saphelp_nw04/helpdata/en/d5/e80d3dbd82f72ce10000000a114084/frameset.htm)
If I design a process chain in which I delete the overlapping request first and then build index then in the checking view it does not give me any error.
The process chain also works fine.
How does it hurt to have it this way or what is the concept behind having the sequence as recommended by SAP.
Thanks,
sam

Hi Sam,
Writing performance is increased when there is no index --- So we delete the index before updating data.
Reading Performance is increased when there is a index -- So we create Index so that when queries are executed quickly by reading the data in the info-cube.
BI Best solutions suggests that old requests are to be deleted before loading new one.So we delete ovelapping requests.
Hope it helps.

Local Process Chains

HI Guys,
I am in the process of designing the process chains for data loading. I have a situation where I am planning to use local process chains in the meta chain. Situation is like this:I have created Meta Chain for Master Data loading, which is having 2 local process chains, one is for Attributes and another is for loading texts. I placed them in the sequential order. First Attributes chain and on successful completion of Attribute chain it should run the Text process chain. But in the system it executes the Attribute chain successfully but the text process chain is always in the YELLOW( in progress ) state even though in that process chain I have only start process. In the SM50 it shows two background process with the user id ALEREMOTE, but process chain never finishes. In the Messages of the Text process chain I see the message"Communication buffer delted from the previous run ".
So I put the process chains in the parallel manner, meaning after start process of Meta chain I attached the two process chains so that they should execute simultaneously. But in this case too, Attribute local process chain gets executed but not the Text local process chain.
We tried debugging the process in SM50, it seems that there is some infinite loop.
If anybody has any clue then please reply back at your earliest. Your help is highly appreicated,
Thanks.
Santosh Taware

first point, i think you can have single process chian for both attributes and texts one after other.
second thing check in st22 if any dumps are occuring and check in the monitor in which step it is still running.
reward if helps.

How to handle hierarchies in process chain

Hi All,
I am designing the process chain for a master GL account number . It has attribute hierarchy and text.
I am successfully able to create the process chain for the attribute and text . But when i add the infopackage for the hierachy I am confused as I have two hierachy infopackages as the client is using two hierachies for the GL accounts.
How should i incorporate these two infopackages which are refereing to two different hierarchies of R3.
I will appreciate your replies.
Regards
Rahul Bindroo

Hi all,
But when I try to put the infopackage with first hierarchy in the process chain I dont get the save hiearchy and attribute change run automatically while as for attributes i get the attribute change run automatically.
If I have to add the load process 'Save hierachy ' i do the following steps:--
Drag and drop save hiearchy--->if i use F4 no list appears/if i create the variant i get a screen with the object type field -
What should I select there.......?>**
Am I following right steps or I am missing something.
Regards
Rahul Bindroo

Process Chains (Urgent)

Hi,
Can any one pls tell me how to create process chains,as i am new to process chains.
Pls give me the steps how to design the process chains.
Thanks in Advance.

hi,
try
http://help.sap.com/saphelp_nw04/helpdata/en/67/13843b74f7be0fe10000000a114084/frameset.htm
https://www.sdn.sap.com/irj/servlet/prt/portal/prtroot/com.sap.km.cm.docs/library/events/sap-teched-03/using process chains in sap business information warehouse
https://www.sdn.sap.com/irj/servlet/prt/portal/prtroot/docs/library/uuid/19683495-0501-0010-4381-b31db6ece1e9

Process Chain Assining for another source sytem

Hi Experts,
I have a quick question on process chain assining for another source sytem.
For Example. I have a Process chain with 10 Infiopackages. Letz say that Infopackage was assinged to one source system.
Now I wanna a create a new process chain which shoud be the replica of the existing process chain. I wanna to load the data for the same info packages thru another sources system. Can we do this or else we have to reassign the data source for new source system for every infosources.. Is this Possible.
Kindly guide me if there is any nother solution for this.
Warm Regards,
Aluri

Hi..
ABAP routines are in place at infopack level or at the Transformation level?
If its there at the Transformation level then there wont be any issues for ur redesigning..
if its there at the IP level then u need to copy those things in ur new IP as per the requirement and change the things accordingly..
and then design the process chain..
Assign points if this helps..
Thanks

Customer Master Process Chain

Hi All,
I am designing a process chain to load my Customer Master Data. I am planning to design as:
Cust. Attr(D) -> Cust. Text(D) -> Cust. Comp code attr(D) -> Cust. CompCode Text(F) - Cust. Sales Attr(D) -> Cust. Sales Text(F).
My questions are:
1. Will this chain takes time becoz of the full loads for Cust. Company code and sales view texts?
2. Do I need to use seperate attr. change run processes for customer, cust. compc view and sales view or 1 for all is enough?
3. Do we need to do the full loads for the texts daily or is it enough to do them periodically?
4. Is there any other better way to design it?
Any inputs will be highly appreciated!!!
Rgds,
Nene.

answers...
1..records for these text loads..may not be high in number..shouldnt take much time..
2..use just 1 attr/change run process..in the variant of the change run process type..do remember to specify these 3 MD infoobjects..
3...this answer depends on whether texts change with regular frequency..like daily..usually texts dont change so regularly..usually placed as weekly loads..
4..maybe this design is better..?
start variant
then..
Cust. Attr(D) --> Cust. Comp code attr(D) -> - Cust. Sales Attr(D)
(3 in parallel..)
Cust. Text(D)Cust. CompCode Text(F)Cust. Sales Text(F).
(3 in parallel..connect respective attr load to its text load..by GREEN arrow)
attr/change run..(connect GREEN arrows from the 3 text loads..to the change run process)..
cheers,
Vishvesh

Design of process chain

Similar Messages

Maybe you are looking for