cancel
Showing results for 
Search instead for 
Did you mean: 

BPC NW - Questions on How to Automate Master Data Loads

Former Member
0 Kudos

Hello All,

We recently released a How to for automating master data loads from BW to BPC (Can be found [Here|https://www.sdn.sap.com/irj/scn/go/portal/prtroot/docs/library/uuid/00380440-010b-2c10-70a1-e0b431255827]). Please post any questions/comments related to this how to guide in this thread.

We just wanted to consolidate all questions and answers under this thread.

Thanks,

Velavan.

Accepted Solutions (0)

Answers (45)

Answers (45)

Former Member
0 Kudos

I am using "How To... Import Master Data and Hierarchies into SAP BusinessObjects BPC 7.5 from SAP NetWeaver BW"

I am using the Scenario 1: "BPC Cost Center dimension member ID = the BW Cost Center number"

Now I am working in the transformation file.

*OPTIONS

FORMAT = DELIMITED

HEADER = YES

DELIMITER = TAB

*MAPPING

ID=ID

COMPANYCODE=0COMP_CODE

*CONVERSION

ID=CeCo_ID_MasterData.xls

My conversion file:

EXTERNAL INTERNAL

* js:%external%.toString().replace(/s+/g,"")

But when I validate the transformation file it shows an "error: all records are rejected".

"No master data for source infoObject 0COSTCENTER Exists"

In the options from "validate and process transformation" I selected the following:

DATA TYPE = Master Data/Text from NW BW InfoObjects

INFOOBJECT=Cost Center

SET SELECTION:

Hierarchy=BPC-10

Version=empty

Member ID = ESSABPC-10

Level = 4

Language=Spanish

Attribute List = Controlling Area

FORMAT=External Format

SELECT A DIMENSION= CECO

¿What do I have to do?

I have search in all documents, but I cannot solve it.

Former Member
0 Kudos

Hi Velevan,

Actually we did install BPC recenty in our BW Sandbox and currently we are trying to load the Master data including hierarchy for BPC infoobjects from BW. I did follow the scenario 2 in 'How to' guide to upload the hierarchy for BPC infoobject Profit_ctr.

After loading Master data attributes and texts using DTP, I did run the program ZUJA_UPDATE_BPC_DIM for a selected Hiearchy under BW infoobject 0Profit_ctr. Program ran successfully and created a dummy node in hierarchy PARENTH1 for BPC infoobject Profit_ctr.

As you suggested, I did make changes in FILL_PARENT_DATA and FILL_TEXT_NODES methods in class ZCL_UJA_DIM_BW_ETL as per page 5 on this thread. Unfortunately it didn't resolve my problem. I'm still seeing a dummy node in PARENTH1 hierarchy.

Could you please let me know the cause for this problem?

Thanks,

Krish

Former Member
0 Kudos

Hi,

Never mind. I found the solution. The problem is that apparently the prompt file should not contain the '& l t ;' and '& g t ;' (without spaces) which are used in the log file that you can download from UJFS. Instead < and > should be used.

Now the loading works fine.

Kind regards,

Christophe

Former Member
0 Kudos

Hi,

We are implementing BPC 7.5 for NetWeaver. As suggested in a blog set up by Pravin Datar (/people/pravin.datar/blog/2009/08/28/automate-execution-of-any-data-manager-package) and as mentioned in your document I am trying to automate Master data and transaction data loading from BI 7 towards BPC 7.5 for Netweaver by means of program UJD_TEST_PACKAGE. But it dos not work for the moment.

My prompt file is defined as follows (taken as a copy of the fixed part of the logfile from a successfull manual execution of the DataManager Package):

%INFOOBJECT% BPCPRCTR

%SELECTION% <Selections><Selection Type="Attribute"/><Selection Type="Hierarchy" ImportText="1"/><Selection Type="Language" TextType="3"><Language>E</Language></Selection><Selection Type="AttributeList"><Attribute>BPCPRCTR</Attribute><Attribute>0CO_AREA</Attribute><Attribute>0RESP_PERS</Attribute></Selection><KeyDate>20100802</KeyDate><Relation><Value Left="Attribute" Right="Hierarchy">1</Value></Relation></Selections>

%WRITEMODE% 2

%FORMAT% 0

%TRANSFORMATION% \ROOT\WEBFOLDERS\ZTEST_KONE\ASSETS\DATAMANAGER\TRANSFORMATIONFILES\BW TRANSFORMATION FILES\BPCPRCTR Transformation.xls

%DIMNAME% PROFCENT

The program executes fine from the process chain each time, but the log in the DataManager allways returns with the following error and no date is loaded:

An exception with the type CX_ST_MATCH_ELEMENT occurred, but was neither handled locally, nor declared in a RAISING clause

System expected element 'Selections'

Do you see any problems in the prompt file above? Or maybe other experience with these error?

Thanks in advance,

Christophe Posson

Former Member
0 Kudos

Has any client managed to load master data hierarchies in BPC NW 7.5 where the master data is compounded with a numeric value Controlling Area e.g. 1000.

We are following the NW BPC 7.5 How to Guide on loading master data and implementing scenario 3 e.g. Cost Element is being prefixed u201CCELu201D.

Loading the cost element attributes works fine with the 7.5 new data package and the hierarchy members/text nodes are correctly populated in BPC.

In our transformation file, it is not possible to use SUPPRESSCHARACTER=1000 as the Controlling Area = 1000 and additionally there are cost elements which contain u20181000u2019 ( and cost centres, profit centres etc)

In the hierarchy transformation file, weu2019ve tried numerous permutations of trying to suppress u20181000u2019 for where the node is a hierarchy text node ( see example below), however it appears that the system just ignores any rules in the transformation file to ignore value u20181000u2019.

NODENAME=IF(NODENAME(1:4)=STR(1000) then NODENAME(5:14);*STR(CEL_)+NODENAME)

As mentioned we cannot use the SUPPRESSCHARACTER statement and master data attributes works fine.

We need to resolve this as it also impacts cost centre and profit centre

Jeffrey_Holdema
Active Contributor
0 Kudos

The syntax you shared does seem OK so I'm not sure exactly what is the issue. Can you provide the error message? Also an update the BPC 7.5 NW [HTG|http://www.sdn.sap.com/irj/scn/index?rid=/library/uuid/c02d0b47-3e54-2d10-0eb7-d722b633b72e] has been posted which separates scenario 3 into two variations. Perhaps 3A will will work for you since it does not use SUPPRESSCHARACTER.

Thanks and best regards,

[Jeffrey Holdeman|http://wiki.sdn.sap.com/wiki/display/profile/Jeffrey+Holdeman]

SAP Labs, LLC

BusinessObjects Division

Americas Applications Regional Implementation Group (RIG)

Former Member
0 Kudos

BI Programme,

Did you get a solution to this problem?

I am interested to hear your outcome as I am in the same situation.

Thanks

Former Member
0 Kudos

Hi Velavan,

I have uploaded the two files given in document, into my BIW system.

I am getting some error as follows:

1. The process chain got save activate with errors. I am working on this.

2. Some ABAP errors like passing parameter type conflict to class are occuring.

The attribute DO_CONFIG is having type ref to IF_UJD_CONFIG in your defined class. Hoever, it gives error while activating the class as the method GET_PARAM_WITHLOG for class CL_UJD_UTILITY having diffrent asociate type for importing parameter IO_CONFIG.

In some places CL_UJD_CONFIG is suitable type for DO_CONFIG and in some cases CL_UJD_PIPE_SIMPLE_EXECUTE is suitable.

It creates somewhat messy situation and results into DUMP.

Can you help on this?

regards,

Sachin

danielpurucker
Participant
0 Kudos

In BPC 7.5 there are several Data Management Packages like "Import BW MasterData" or "Import BW Master Data Text". How do these substitute this Howto? Or has this document still advandaces (despite it can load delta records)?

Regarding the point of Christoffer Owe: Are there any updated transports available?

thanks for sheding light on this.

danielpurucker
Participant
0 Kudos

Anybody?

danielpurucker
Participant
0 Kudos

Accoring to the attached thread this howto is obolete.

This thread also describes the error message, when importing the "old" attached transports.

Jeffrey_Holdema
Active Contributor
0 Kudos

Indeed the BPC 7.0 HTG on How to Automate Master Data Loads is NOT applicable for BPC 7.5 NW. There are now standard Data Manager process chains for loading BW master data and hierarchies in BPC 7.5 NW. See the following [HTG|http://www.sdn.sap.com/irj/scn/index?rid=/library/uuid/c02d0b47-3e54-2d10-0eb7-d722b633b72e].

Thanks and best regards,

[Jeffrey Holdeman|http://wiki.sdn.sap.com/wiki/display/profile/Jeffrey+Holdeman]

SAP Labs, LLC

BusinessObjects Division

Americas Applications Regional Implementation Group (RIG)

0 Kudos

Hi gurus,

I am facing the follow problem: after I export .csv , I check the description and the hierarchy in the file .cvs and its ok, but when i execute the package to load master data, this 2 fields donu2019t load.

My mapping:

ID=*COL(1)

CO_AREA=*COL(2)

COSTCENTER=*COL(3)

DATETO=*COL(4)

PROFIT_CTR=*COL(5)

EVDESCRIPTION=*COL(6)

PARENTH1=*COL(7)

Anyone having this problem?

Thanks

Former Member
0 Kudos

Hi All,

While loading master data from BW to BPC using scenario 2 of the "How To Automate BPC 7.0,version for the

NetWeaver Platform Master Data Loads from BW to BPC Version 1.50" we are facing a strange error.

The data load is successful and we are able to see the dimension members in BPC frontend.

When we do a transactional data load in the application (through data manager from SAP BW cube) of which this dimension is part of is also workking ok.

The data is also available in the BW backend in the BPC cube but we see no data in the EVDRE report.

This is not happening once but every time when we automate the masterdata load from SAP BW infoobject to BPC dimension the transactional data disappears.

regards,

Atul

Former Member
0 Kudos

Hi,

We ZUJA_UPDATE_BPC_DIM as part of a daily scheduled process chain to copy the hierachy from infoobject 0ACCOUNT to P_ACCT. From BPC, users modify the node's and leaf's attribute ACCTYPE. When the ZUJA_UPDATE_BPC_DIM is ran again, the user's hierarchy node's ACCTYP modifications are wiped out. However, the leaf ACCTYP modification is still there. Is there any way to keep the ZUJA_UPDATE_BPC_DIM from the modifying to the nodes' attribute data?

Thanks.

Regards,

Mel Calucin

Bentley System

Former Member
0 Kudos

Hi Mel,

I had something similar. I solved it by putting an attribute change run for the BPC dimension InfoObject into my process chain. The change run needs to be placed after the DTPs but before the program ZUJA_UPDATE_BPC_DIM.

Reason is that some BPC dimensions are created as NAV attributes. If they change, you need to activate your master data before you see them anywhere. The change run does that. Above mentioned program seems to ignore this.

Hope that helps,

Steffen

Former Member
0 Kudos

Hi,

I implemented scenario 2 of the HTG and discovered the following problem:

After the masterdata load and attribute change run for 0costcenter, all values are visible in the masterdata table and all HIR fields are empty.

Then I start the program ZUJA_UPDATE_BPC_DIM to read my hierarchy (there's only one).

After a successful run, all values of 0costenter are assign to hierarchy H1, although they are not all used in this hierarchy.

Does anyone has an idea how this is possible?

Kind Regards,

Thierry

Former Member
0 Kudos

Hi,

are the values assigned to different levels and/or nodes of the hierarchy or only below top node?

Are intervals used within the hierarchy in BW?

Regards

Steve

Former Member
0 Kudos

Hi David,

Did you ever get this issue resolved? If so, can you share the solution? I am running into the same issue of the cost centers all being assigned to H1 instead of following the ECC hierarchy structure.

I'm thinking it might be a combination of how the calculated member field is populated, and the fact that the text nodes are not dimension members. We had already created a hierarchy in the BPC dimension and I noticed when I maintain the cost center master data the "parents" in the BPC dimension show calculated member = Y.

Any, and all, advice is appreciated.

Thanks,

Pam

Former Member
0 Kudos

Hi,

I implemented scenario 2 of the HTG and discovered the following problem:

After the masterdata load and attribute change run I c

Former Member
0 Kudos

Hi,

I implemented scenario 2 of this howto.

After a successful run, the hierarchy is created but I only see a dummy node.

No other node or leaves are visible, although the source hierachy is correct and active.

Does anyone know the cause for this?

Kind regards,

Thierry

Former Member
0 Kudos

I got this aswell, numerous times. It seems to be an indicator that something went wrong in the 'Process Dimension' step of the program, and it just leaves you with a mess, but does not return an error message.

Try to get your dimension with the hierarchies into a format so that you can upload it manually via the Admin client. Then process the dimension and wait to see what errors you get.

What you might have is two nodes that share the same technical name in two hierarchies on the dimension, where BPC tries to create 2 members with the same technical name. In this case, you will need to change the hierarchy node name in one of the dimensions.

Alternatively, there may already be an issue with your dimension - just try processing the dimension without the hierarchy and see if you get an error.

Hope this helps,

Peter

Former Member
0 Kudos

I had similar issue and upon checking the program ZUJA_UPDATE_BPC_DIM logic, the program does verify the validity of 'A_ACCTYPE' or ACCTYPE within Infoobject P_ACCT against a set of valid values which were 'AST', 'EXP', 'INC' and 'LEQ'.

After modifying the transformation and loaded the valid values for ACCTYPE, then that program successfully constructed the hierarchy.

regards,

Indra

Former Member
0 Kudos

Hi,

I'd like to share the code tweaking that allows us to construct BW hierarchy that uses interval. The intervals are now translated into actual nodes under their respective parent.

The following codes are inserted within method 'FILL_INTERVALS' of CLASS 'ZCL_UJA_DIM_BW_ETL'.

Insert these logic right after the the last 'LOOP' and before the statement 'e_success = abap_true'.

Regards,

Indra

....

REFRESH <htable>.

ENDLOOP.

*{ INSERT

DATA: l_hieint type string,

l_hietab type string,

wa_nodeid type rshieid,

wa_nodenm type rsshnodename.

  • -------- Check and complete PARENT ID for interval nodes------------------

  • Store hierarchy interval table in internal memory

CLEAR: l_hieint, l_hietab, wa_nodeid, wa_nodenm.

temp = 'PARENTH'.

  • --- Determine the source hierarchy tables (incl. interval table).

If l_fieldnm(1) = 'Z'.

CONCATENATE '/BIC/J' l_fieldnm INTO l_hieint. "Interval

CONCATENATE '/BIC/H' l_fieldnm INTO l_hietab.

Else.

CONCATENATE '/BI0/J' l_fieldnm INTO l_hieint. "Interval

CONCATENATE '/BI0/H' l_fieldnm INTO l_hietab.

Endif.

LOOP AT it_data ASSIGNING <ls_line>.

LOOP AT nt_hierarchies INTO ls_hierarchies.

l_index = sy-tabix.

CONCATENATE temp l_index INTO parenth.

ASSIGN COMPONENT parenth OF STRUCTURE <ls_line> TO <parenth>.

ASSIGN COMPONENT 'ID' OF STRUCTURE <ls_line> TO <tvalue>.

ASSIGN ls_hierarchies-r_hierstruc->* TO <r_hierstruc>.

LOOP AT <r_hierstruc> INTO ls_hierstruc FROM 1 TO 1. "Get the hierarchy ID

ENDLOOP.

IF <parenth> is INITIAL. "Process for unassigned parent node

SELECT SINGLE NODEID FROM (l_hieint) INTO wa_nodeid

WHERE HIEID EQ ls_hierstruc-hieid

AND leaffrom le <tvalue>

AND leafto ge <tvalue>.

if sy-subrc eq 0. "Get the parent node ID

SELECT SINGLE NODENAME FROM (l_hietab) INTO wa_nodenm

WHERE HIEID EQ ls_hierstruc-hieid

AND NODEID eq wa_nodeid.

IF SY-SUBRC EQ 0.

<parenth> = wa_nodenm.

ENDIF.

Endif.

Endif.

ENDLOOP.

ENDLOOP.

*} INSERT

e_success = abap_true.

Former Member
0 Kudos

Hi Indra,

I'm attempting to add your code and am receiving a syntax error stating that "parenth" doesn't exist but that there is a similar element <parenth>. Did y'all have to go back and declare this element?

Thanks,

Pam

Former Member
0 Kudos

Hi Pam,

I think you missed number 1 behind PARENTH because it should be PARENTH1.

Ba

Former Member
0 Kudos

Hi Ba,

This is an excerpt from the code. Where are you thinking the parenth1 should be placed? The original code in the method has the following declaration at the top:

temp TYPE string VALUE 'PARENTH'

===========================================================================

temp = 'PARENTH'.

  • --- Determine the source hierarchy tables (incl. interval table).

If l_fieldnm(1) = 'Z'.

CONCATENATE '/BIC/J' l_fieldnm INTO l_hieint. "Interval

CONCATENATE '/BIC/H' l_fieldnm INTO l_hietab.

Else.

CONCATENATE '/BI0/J' l_fieldnm INTO l_hieint. "Interval

CONCATENATE '/BI0/H' l_fieldnm INTO l_hietab.

Endif.

LOOP AT it_data ASSIGNING <ls_line>.

LOOP AT nt_hierarchies INTO ls_hierarchies.

l_index = sy-tabix.

CONCATENATE temp l_index INTO parenth.

ASSIGN COMPONENT parenth OF STRUCTURE <ls_line> TO <parenth>.

ASSIGN COMPONENT 'ID' OF STRUCTURE <ls_line> TO <tvalue>.

ASSIGN ls_hierarchies-r_hierstruc->* TO <r_hierstruc>.

LOOP AT <r_hierstruc> INTO ls_hierstruc FROM 1 TO 1. "Get the hierarchy ID

ENDLOOP.

Thanks,

Pam

Former Member
0 Kudos

Hi Pam,

Pardon me, never mind about what I wrote earlier because I did not pay attention to that extra code. For me after some upgrades the hierarchy interval worked for scenario 1; therefore, I did not use that code.

Regards,

Ba

Former Member
0 Kudos

Hi Thierry,

Were you able to resolve this problem (Hierarchy is showing as Dummy Node)?

I'm also having the same problem after running a program 'ZUJA_UPDATE_BPC_DIM' to upload hierarchy.

Though program ran successfully, when I check the loaded hierarchy for 0Profit_ctr, it shows as Dummy node even though my source hierarchy is active and correct.

Could you please share the steps you have taken to fix this problem?

Thanks,

Krish

Former Member
0 Kudos

Hi Krish,

Exactly when/where are you seeing this issue? We experienced the Dummy Node after transporting the program and process chain up to our production environment. It turned out our source hierarchy was active in dev and qa, but not in production. Once we created the infopackage and loaded that particular hierarchy in production, and executed the program, the hierarchy was visible.

We also experienced the problem Thierry described when first implementing scenario 2 of the HTG. On page 5 of this thread are instructions for editting the FILL_PARENT_DATA and FILL_TEXT_NODES methods in class ZCL_UJA_DIM_BW_ETL. I implemented those changes and the hierarchies loaded correctly (0PROFIT_CTR and 0COSTCENTER).

Hope one of these suggestions helps with your situation.

Regards,

Pam

Former Member
0 Kudos

Hi Pam,

Thanks for your response on this.

Actually we did install BPC recenty in our BW Sandbox and currently we are trying to load the Master data including hierarchy for BPC infoobjects from BW. I did follow the scenario 2 in 'How to' guide to upload the hierarchy for BPC infoobject Profit_ctr.

After loading Master data attributes and texts using DTP, I did run the program ZUJA_UPDATE_BPC_DIM for a selected Hiearchy under BW infoobject 0Profit_ctr. Program ran successfully and created a dummy node in hierarchy PARENTH1 for BPC infoobject Profit_ctr.

As you suggested, I did make changes in FILL_PARENT_DATA and FILL_TEXT_NODES methods in class ZCL_UJA_DIM_BW_ETL as per page 5 on this thread. Unfortunately it didn't resolve my problem. I'm still seeing a dummy node in PARENTH1 hierarchy.

I'm sure there is nothing wrong with source hierarchy. I did try to debug the code and seems like method UPDATE_BPC_DIM is causing a problem here. I'll continue to debug further and see whether I can find the cause for this problem. Would it be possible for you to share the exact changes you made in the program/methods to load the hierarchy for 0Profit_ctr in a successfull manner? I mean using scenario 2.

Thanks in advance for your help.

Bregards,

Krish

Former Member
0 Kudos

Hi Krish,

I didn't do anymore for the methods than what was shown earlier in the thread. As I stated, we saw this problem when the hierarchy we wanted to load wasn't actually set up to load the BW infoobject. Check the infopackages that load the 0PROFIT_CTR hierarchy and verify that the hierarchy you expect to see in the dimension is actually being loaded in BW. The incorrect OLTP hierarchy could have been selected in the infopackage.

Pam

Former Member
0 Kudos

Well, the hierarchy at BW side looks fine and I did verify the infopackage too and found no issues. Still for some reason the program fills this hierarchy node as 'Dummy Node' at BPC side.

Any other thoughts from everyone in this thread?

Thanks,

Krish

Former Member
0 Kudos

Hi folks,

I'm following the 'how to' and i have some problems.

Transport:

The import of the transport ended with RC 8.

In spite of this I tried to continue but I has the following problems:

Scenario 1:

the import data process is cancelled. in Tcode ST22 we saw a dump in class ZCL_UJD_MASTER_LOAD 'ME->DO_CONFIG' is not type-compatible with formal parameter 'IO_CONFIG''

Scenario 2:

In step '4.3.2 Create ABAP program variants' when we try to execute both programs ( ZUJA_UPDATE_BPC_DIM_LOCK and ZUJA_UPDATE_BPC_DIM_LOCK) there is a sytax error.

'Structures with components that are not character-type cannot be displayed on the selection screen.'

thanks in advance.

Carlos

former_member201314
Participant
0 Kudos

Indra,my friend!

Just read this blog. We're facing exactly this issue around the hierarchy intervals and the fact that BPC cannot cater for them when importing.

So maybe your code is about to make it around the world. (if not already happened )

Hope you are traveling well, mate!

All the best from Germany

Claus

Former Member
0 Kudos

Hi,

I implemented the upload of master data, loading etc works fine.

Now I have to change the dimension for reporting purposes. (add some new properties)

When trying to Modify Dimension, I always recieve the following error message:

Failed to change BI InfoObject for dimension

Table UJA_DIM_ATTR shows the correct entries, but program UJA_DATA_CHECKER tells me, that no characteristic was created and therefore could not be added as an attribute.

The program seems to be right, as I can't find it either.

Does anyone know, how to solve this problem?

Modify / Process dimension did not work. (in various sequences)

When copying the dimension, the new dimension has all properties, as defined.

Hoping for answers

Steve

Edit:

I removed transformation, modified the dimension in Admin Console and recreated the transformation.

Everything works fine so far.

Edited by: Steve Scheidig on Mar 23, 2010 1:54 PM

Former Member
0 Kudos

Hi,

just for your information. (I am not sure if it was mentioned here before)

I had the issue that nodes of a loaded hierarchy appear in a different order than created in R/3 or BW system.

To resolve this issue, I removed line 72 in method FILL_TEXT_NODES of class ZCL_UJA_DIM_BW_ETL.

     SORT it_data.

Now the nodes are displayed as defined.

To change the order of an already loaded hierarchy I had to remove all created members (hier nodes) and also the entries in PARENTHx property and load the hierarchy again. (do not forget to save to server and process dimension)

All nodes should were created again without changing the defined order.

I hope it helps. If you see any problems with this procedure, please inform me.

Regards

Steve

Edited by: Steve Scheidig on Mar 18, 2010 8:52 AM

Former Member
0 Kudos

Thanks for the tip on sorting the nodes.

I wanted to share something that I have discovered with the process.

Sometimes the program ZUJA_UPDATE_BPC_DIM will appear to have run correctly, but the hierarchies will not be correct. They may be jumbled, missing members, or plain silly (two hierarchies rolled into one, hierarchies consisting of a single node, etc.).

First thing to check is if the dimension can be processed in the Admin workbench. If not, then sort that out first before trying to load hierarchies.

The reason that I found in my case through debugging was that I had two hierarchies which shared technical node names at some levels in the hierarchy, e.g.

Hierarchy 1:

CORP

->USA->MINNESOTA->Member2

->EUROPE->GERMANY->Member1

Hierarchy 2:

SBU:

->ELECTRONICS->SEMIS->Member2

->PACKAGING->GERMANY->Member4

Hierarchy 1 and Hierarchy 2 share the same technical nodename (GERMANY). I have found that situations like this may cause issues which do not result in an error in program processing, but do result in garbled hierarchies. What is especially vexing here is that the resultant hierarchies may change from run to run - the errors do not appear consistent.

I have found that enhancing the hierarchy extract on ECC using the BADI RSU5_SAPI_BADI gives you an opportunity to rename the technical nodenames by hierarchy (changing C_T_HIENODE-NODENAME).

If in doubt, check the hierarchy table of your source info object for duplicated nodes.

The BADI solved my issue, as I did not want to tell the business users to rename their hierarchies and it is unreasonable to ask them to check to make sure they do not use the same technical node name twice in different hierarchies.

Oh, and another point - if you do get those garbled hierarchies into the system, you need to go into the Admin workbench and delete out the members that have been erroneously created for hierarchy nodes, or BPC will remain confused, but still valiantly chug on and ultimately spit out some jumbled nonsense. Start with a clean slate each time.

Regards

Peter

Former Member
0 Kudos

Hello,

I am trying to extract data from BW system in order to load the same in BPC via Open HUB destination. I have created a transformation for the master data attributes. In the transformation, I have followed the instructions from the sap document i.e. How to automate master data loads, by adding the rule for the destination column EVDESCRIPTION and giving the short text technical name (SHORT/0TXTSH) in the description field of the rule. However, when I run the DTP, I don't get the short description from the info object in the CSV file.

Can some one help me out as to what I am doing wrong?

Regards

Jaideep

Former Member
0 Kudos

Hi Jaideep,

EVDESCRIPTION is updated from the texts you loaded for the characteristic.

That means you have to create a transformation to load texts (in your case short texts) into the text table of the characteristic.

If you then follow the process as described in the documents, your EVDESCRIPTION property should be filled accordingly.

Let us know, if I am right.

Regards

Steve

Former Member
0 Kudos

Dear experts,

Instead of outputing via Open Hub. We wrote a small abap program which will extract from material table and write to a flat file in local (or network) folder. It's been tested that the flat file can work for the MasterDataLoad package by uploading the file to UJFS by hand. How can I automate the MasterDataLoad from here if the flat file will remain in the local (network folder)?

Any suggestion is appreciated,

Many thanks and best regards,

Jim Hsu

federico_blarasin
Active Participant
0 Kudos

I'm implementing Scenario 1 (open hub data loading).

I have a strange issue with a Open Hub.

I'm extracting data from 0COORDER (about 110.000 records in 1 package).

No problems at all in our dev/quality system, extracting into presentation or application server. We used it in our production system without errors at first (saving file into presentation server): now the DTP has got errors on the end routine, raising exception (RS_EXCEPTION 000) with message 'The Database colummn 0DATETO is unknown' (COORDER has not time dependent master data, texts nor hierarchies). Note 1239269 is yet implemented in our system.

There's no custom code in our start/end routine, only what described in 'How to automate master data loading' pdf.

Former Member
0 Kudos

Hi,

I'm trying to implement scenario 2 of this Howto. Thanks for all the work, btw.

It works fine, except there are some issues.

1) When I upload master data for the first time, no problem.

When I do it again after some changes in MD , these changes are not taken into account.

- loading data into characteristic (application set dimension) -> new attributes are visible (in RSA1)

- starting program ZUJA_UPDATE_BPC_DIM with variant -> old attributes are visible (in RSA1 and BPC Administration)

Any suggestions on that?

2) There seems to be a problem with hierarchy update. I started debugging class ZCL_UJA_DIM_BW_ETL:

Method FILL_INTERVALS Line 70f


      CONCATENATE where 'and DATETO =' l_date INTO where SEPARATED BY space.
      SELECT * FROM (n_tabname) INTO TABLE <mdtable> WHERE (where).

If a characteristic has no time dependent attributes or texts there is no 'DATETO' im M-table. Therefore the WHERE-clause is not correct.

This leads to an exception, that is caught, but not processed further

Method UPDATE_BPC_DIM Line 69ff


    CATCH cx_root INTO lo_exception.
      no_dim->release_writelock( ).
      CALL FUNCTION 'UJ0_CONV_EX2MSG'
        EXPORTING
          io_exception     = lo_exception
        IMPORTING
          et_message_lines = et_message.
      e_success = abap_false.
  ENDTRY.
  e_success = abap_true.
ENDMETHOD.

Method should end without setting e_success = abap_true!

In this case, the program ZUJA_UPDATE_BPC_DIM ends with wrong (error) messages.

Maybe this helps to solve some problems, some of you might have.

Regards

Steve

Edit:

Found another issue...

If a hierarchy contains intervals at a certain level, the members of the intervals are not supplied with parent-ID.

Also hierarchies from compounded characteristics are not correct processed. (compounded values are ID for hier-nodes not plain value of characteristic as assumed in method FILL_PARENT_DATA lines 21ff)

I will post a solution as soon as I found a suitable one.

As a remark, in my scenario I did not concatenate characteristics to create a new key, but I selected 1 value of the compounding characteristic...

Edited by: Steve Scheidig on Jan 13, 2010 4:03 PM

Former Member
0 Kudos

Hi,

we have a problem with the how to guide to upload master data from BW. Everything is working except the upload of BW heirarchies. This works fine if the hierarchy is defined with single nodes below each other. However when using intervals in the hierarchy it is not working. In the how to guide under scenario 1 this is explictely mentioned that this not possible, however we are using scenario 2. Also in the example of this scenario an example is dislayed using intervals ...

anybody experience the same issue

D?

Former Member
0 Kudos

Hi Dries,

try debugging method FILL_INTERVALS of class ZCL_UJA_DIM_BW_ETL to check, where errors occur.

1 possibility is time dependency. The method is reading master data tables with 'DATETO=99991231' (line 70)

Another source of errors may be compounding. In my opinion methods GET_IOBJ_INFO and IOBJNM_TO_FIELDNM need to be corrected.

Regards

Steve

Former Member
0 Kudos

Hi Steve,

We checked ...

First of all apperantly the info object needs to have attributes (/BICMxxxxx) table. otherwise it is not working. So we took another one which has a view defined. Apperantly the problem is located here:

LOOP AT nt_hierarchies INTO ls_hierarchies.

l_index = sy-tabix.

<htable> = it_data.

ASSIGN ls_hierarchies-r_hierstruc->* TO <r_hierstruc>.

ASSIGN ls_hierarchies-r_hierintvl->* TO <r_hierintvl>.

LOOP AT <r_hierstruc> INTO ls_hierstruc WHERE iobjnm EQ '0HIER_NODE' AND intervl EQ rs_c_true.

READ TABLE <r_hierintvl> INTO ls_hierintvl WITH KEY nodeid = ls_hierstruc-nodeid.

<md_line> = ls_hierintvl-leaffrom.

In this loop the system checks on iobjnm EQ '0HIER_NODE' , however in our BW hierarchy the interval has the name of the info-object itself instead of '0HIER_NODE' for the interval line ...

=> so the question is how do you achieve this when creating a BW hierarchy? when we add an interval (with the button create interval) the system automatically creates IOBJNR = name of the info object instead of 0HIER_NODE

D

Former Member
0 Kudos

Hi Dries,

Velavan and I have been meaning to publish a new guide for sometime however we have not gotten the time to update everything. I am attaching a transport with the latest and greatest updates here to the forum for your and the rest of the communities benefit.

There are some things to watch out for. Specifically if you are using scenario 1 of the guide and you are trying to load from an infoObject into a BPC Dimension where the infoObjects technical length is greater than 20 (the maximum length for a BPC Dimension) it automatically generates a hash key for the ID value that is 20 characters in length. This is not the behavior when running scenario 2.

I hope this update helps.

Cheers, Scott

Former Member
0 Kudos

Hello,

We have implemented scenario 1 and we are encountering issues with the language of the description. Everything should be transferred in Dutch but during the automatic update it is transferred in English.

As our BPC clients are in Dutch the description doesn't appear at all in our reports. What we should do is a copy in the text tables to the "NL" language...

When I am doing a manual update in the BPC admin the description is transferred with the correct language.

Has someone an idea on how we can change the language of the description during the automatic update .

Thanks,

Mathieu

Former Member
0 Kudos

Hi,

if I remember correctly the transferred language is dependent on the language setting of the user.

So if you are starting the upload (with dutch settings) dutch descriptions are transferred.

Please check the settings of your standard user, who is doing the automated upload.

regards

Steve

Former Member
0 Kudos

Hi Scott,

we imported the new files attached in your previous response, but still are experiencing problems:

1) the fill parent member method still did not work: method 'FILL_INTERVALS' of class ZCL_UJA_DIM_BW_ETL : (<line> was not compatible with l_id)

2) the locking program always ended in error: changes to program ZUJA_UPDATE_BPC_DIM_LOCK :

The last if was returning in each case an error-message. I've removed the if-command on l_success and now its given me that the dimensions are unlocked and the program runs succesfully

I can send you the details of the programming by mail (too long to post here ...)

3) sometimes we get error in BPC Admin "could not activate BI info object ..." and a discrepancy occurs between the file in UJFS folder structure and the detailed members in RSD1 ... Only when solving the issue on the BI info object directly (by deleting the property we were trying to create additionally in BPC Admin) the files structure and RSD1 structure are consistent again ... This problem we only get after loading the master data by means of the how to guide procedure ...

D

Former Member
0 Kudos

Hi Scott,

Using scenario 1 of this How-to to extract 0GL_ACCOUNT out of BW but Execution (DTP) failed because 0GL_ACCOUNT has alphanumeric value.

If failed at Method ID line 36 u201Cresult = l_id(n_length).u201D in class ZCL_BPC_MD_AUTOMATION.

ID = 20 characters therefore u201Cresultu201D is defined as 20 characters. 0GL_ACCOUNT with 10 characters and has a compounding object which is 0CHRT_ACCTS with 4 characters so total is 14 characters but some how it failed there.

Please look into this for me. Thanks.

Hi Scott,

I am updating you with the good result.

After using the new transport MDA_-_20100121.zip the length error which I mentioned above went away but the extraction failed. We think it failed when a certain number of type I messages had reached its limit. Therefore, we ignored type I message in method ADD_MESSAGE from class ZCL_BPC_MD_AUTOMATION and the extraction worked.

From below this code in method ADD_METHOD:

assign component 'MSGTY' of structure <line> to <msgty>.

if <msgty> is assigned.

<msgty> = i_type.

endif.

We inserted at line 33

*{ INSERT BWTK903523 1

*NSC changes to delete the messages type "I'

if <MSGTY> NE 'E'.

EXIT.

ENDIF.

*} INSERT

So it's working now.

Regards,

Ba

Edited by: Ba Tran on Feb 16, 2010 11:56 AM

Former Member
0 Kudos

Hi all

I am working on Scenario 1

while automating the load from BW to BPC using the program : UJD_TEST_PACKAGE i have an issue.

But i am able to run the same using the DM package sucessfully. But When i execute the program i have an error called :

does anyone faced this issue;

/CPMB/MODIFY : Completed in 0 seconds

/CPMB/CLEAR : Completed in 0 seconds

[Selection]

-


FILE \usr\sap\DB1\DVEBMGS00\work\ZOH_CC.CSV=

TRANSFORMATION \ROOT\WEBFOLDERS\FINPLANFORECASTING\FORECASTING\DATAMANAGER\TRANSFORMATIONFILES\MD_CC_T.xls=

DIMNAME COSTCENTER=

UPDATE 1=

[Messages]

-


XML file (...%TRANSFORMATION%) is empty or is not found

Cannot find document/directory

Application : FORECASTING. Package status: ERROR

Former Member
0 Kudos

Hi,

You may want to double check the path of the transformation file for correct app set, app, and directory. Also check for different version of Excel if you are using Excel 2003 and Excel 2007.

Regards,

Ba

Former Member
0 Kudos

Thanks Ba,

The issue was with the file path. corrected the file path.

it is resolved now.

Thanks

Former Member
0 Kudos

Hello,

It seems the updated transports are no longer available in the forum? Will you please repost them? We have several issues with the transports in the document from June 2009.

Best regards,

Christoffer Owe

Former Member
0 Kudos

Hi Scott,

It seems like the transport with the latest updates to the program is not available in this tread. Can you please provide us with this?

Thanks

BR,

Linda Solberg

Edited by: Linda Solberg on Aug 19, 2010 11:37 AM

Former Member
0 Kudos

Hi all,

We have used scenario1 and created 4 open hubs in order to load master data from BW to BPC. They are all working fine, except from the one based on 0ACCOUNT. When we execute the dtp we get the error message: "Exceptions in substep: End Routine". When I right-click on the message and choose display messages, I see the following: "The database column 'DATETO' is unknown" (RS_EXCEPTION000).

I've double-checked that all details for the Open Hub, transformation and dtp, and can't see that anything is wrong.

There is no DATETO or DATEFROM attributes on 0ACCOUNT, and I'm not sure why it displays the error message regarding the DATETO.

Does anyone have any suggestions on what to do?

BR,

Linda

Former Member
0 Kudos

Hello experts,

I use the HTG (scenario 2) to load master data into a BPC customer dimension, from 2 standards BW infoobjets : 0cust_sales and 0customer (I have to get records from both, content is not the same in both infoobjects).

I skip steps concerning the time dependance and compounding management (24 to 34), and everything work fine for 0cust_sales : /CPMB/CALC is populated with 'N', after an UJA_REFRESH_DIM_CACHE i can see new dimension members in BPC admin and also in BPC for Excel,

... but for 0customer, I don't have any hierarchy to replace (0customer is only a flat list, no hierarchy is loaded),

In document it is step 10 of part '4.3.2 Create abap program variant' : when I hit 'enter' as I just say since no hierarchy is available I can't choose anyone, I save the variant without hierarchy.

so the loading is successfull, but with a "maintain master data" I can see that HIR and Calc member stay empty, new records are available in BPC Admin but they are not visible in BPC for Excel (and later when I load actual data from BW to BPC, all records concerning those customers are rejected).

I tried to force Hir and Calc member to 'H1' and 'N' with rule constant in transformation, fields are populated in table, but the problem stay the same on BPC side.

I spent a lot of time on that, and any help will be highly appreciated.

Thanks in advance.

Edited : Still nobody on this thread ?

Former Member
0 Kudos

I'm working in a scenario where I automate master data load directly from a BW info object. User on the other hand are allowed to change dimension properties after this master data load.

Problem is when I load master data from BW, table gets updated but not dimension file.

Is there any program that will update (sync) dimension file from dimension table?

Thanks

ps: I've posted already this question on

Former Member
0 Kudos

I've found a solution to my problem. Recaping:

1. Load master data from a BW info object into BPC (scenario 2 on how to document)

2. Data gets loaded directly into BPC tables but I don't see these changes on BPC dimension excel file

I changed method ZCL_UJA_DIM_BW_ETL -> UPDATE_BPC_DIM:

On line 28, instead of:

me -> increase_srv_ver( ).

I used:

nodim -> update_dim_versions( ).

It seems it is not enough to simply increase server_version on table UJA_DIMENSION as INCREASE_SRV_VER does in order to have both table and excel in sync.

Former Member
0 Kudos

We are experiencing error in importing the transport PMRK900733 to implement the scenario mentioned in the automate the dataload. I would appreciate if anyone can point out what might be the reason for this. We are on NW 7.0 SP5.

Thanks.

Former Member
0 Kudos

Hi,

I am getting the following error when trying to run DTP for an Openhup for 0profit_ctr (ATTR, HIER and TEXT) extract:

Runtime Errors ITAB_DUPLICATE_KEY

An entry was to be entered into the table

"" (which should have

had a unique table key (UNIQUE KEY)). However, there already existed a line with an identical key.

Please can you help on this.

Thanks

Gaurav

Former Member
0 Kudos

I'm trying to implement scenario 2 to load our Profitcenter hierarchie into BPC, but unfortunately I can't get it to work properly. The text-nodes aren't placed in the correct levels.

The hierarchy consists of multiple text-nodes (on different levels) with multiple characteristic-nodes underneath.

Below you'll see what happens.

Source hierarchy (before running ZUJA_UPDATE_BPC_DIM):

Textnode1 (level 1)

Textnode2 (level 2)

Textnode3 (level 3)

Characteristic-node1 (level 4)

Characteristic-node2 (level 5)

Characteristic-node2 (level 5)

Destination hierarchy (after running ZUJA_UPDATE_BPC_DIM):

Textnode1 (level 1) <== Changed to characteristic node

Textnode2 (level 1) <== Changed to characteristic node

Textnode3 (level 1) <== Changed to characteristic node

Characteristic-node1 (level 2)

Characteristic-node2 (level 3)

Characteristic-node2 (level 3)

Please advice.

Regards,

Vincent

Former Member
0 Kudos

Hey Vincent

Probably two changes are necessary within the coding of ZUJA_UPDATE_BPC_DIM.

First I have changed the method FILL_PARENT_DATA (class ZCL_UJA_DIM_BW_ETL)

At line 27 I have removed the check on the source infoobject. This makes it possible to select 0HIER_NODE parents as well.

*{   REPLACE                                                1
*\        iobjnm    = n_iobj INTO l_hierstruc TRANSPORTING ALL FIELDS.
        INTO l_hierstruc TRANSPORTING ALL FIELDS.
*}   REPLACE

The second change was done in the method FILL_TEXT_NODES (class ZCL_UJA_DIM_BW_ETL).

This makes it possible to update text nodes instead of only adding new text nodes (at line 68)


*{   REPLACE                                                1
*\        IF sy-subrc NE 0.
*\          INSERT <ls_line> INTO TABLE it_data.
*\        ENDIF.
        IF sy-subrc NE 0.
          INSERT <ls_line> INTO TABLE it_data.
        ELSE.
          MODIFY TABLE it_data FROM <ls_line>.
        ENDIF.
*}   REPLACE

Hope this will solve the issue.

Regards,

Steven Groot

Former Member
0 Kudos

Hi Steven,

We are experiencing the same problem when loading a hierarchy from a BW infoobject to a BPC Dimension: the text nodes are displayed as characteristic node, and the real characteristic nodes are listed without being included in the hierarchy.

We have implemented the correction suggested by you, but this did not solve the problem, we still get the same result as before.

Has anyone else experienced this problem and solved it?

Thanks.

Best regards,

Linda

Former Member
0 Kudos

Has anyone been able to pass in the AnswerPrompt file in for section 3.2.6?

I've created the flat file per the instructions (with a tab between the variable and value and new line for the next), however the code just returns a blank value to the AnswerPrompt input field. I'm not sure if there's a problem with my file or with the WS_UPLOAD function module that is called.

Thanks.

-Amy

Former Member
0 Kudos

Hi all,

i followed 'Master Data Loads from BW to BPC' guide step by step but when i am trying to run 'Import_How_To_Md_Automation' data package , i am facing the following error :

No 1 Round:

An exception with the type CX_SY_STRUCT_COMP_NAME occured, but was neither handled locally, nor declared in a RAISING clause

Component Name 'DTP;DTP_D6T22N9KFRMKZE1V1DTXH2YH4;;;;;' of the Component 16 Too Long

Any ideas about this problem?

Thanks..

Former Member
0 Kudos

Vanya,

Since it has been long time since you posted your query, I am assuming you were able to solve your problem. Could you share with me what was the cause and resolution of this error?

Thanks.

Former Member
0 Kudos

HI

I'm trying to follow the How to "How to Automate Master data Loads" . The first step is to import the 2 .*PMR files with STMS transaction. So I put the 2 file in the directory :

K* file into /usr/sap/trans/cofiles

R* file into /usr/sap/trans/data

and than i run the transaction stms. Here under the menu "extras->other request->add" I'm not able to see the added request (pmr files).

Any ideas what I'm wronging??

Former Member
0 Kudos

The program ZUJA_UPDATE_BPC_DIM doesn't seem to work when I select a hierarchy to load. I receive an error message which doesn't tell much ("HIERARCHY343£%£"). Does anyone have an idea what may cause this ?

Thanks in advance ...

Former Member
0 Kudos

SCENARIO 2

Two small remarks about the update program ZUJA_UPDATE_BPC_DIM :

- When you have defined your Application Set in BPC including lowercase letters (for example "Consolidation"), the program doesn't work. You have to remove the code where the p_appset parameter is converted to uppercase because the application set id in BPC is case-sensitive.

START-OF-SELECTION.

  • TRANSLATE p_appset TO UPPER CASE.

This also applies to the program ZUJA_UPDATE_BPC_DIM_LOCK.

- When you want to load only text masterdata to an infoobject used by BPC, the update program ZUJA_UPDATE_BPC_DIM won't work, because the BPC specific attributes that are updated are in the attribute table and NOT in the text table. So you have to make sure that all the masterdata id's are in the attribute table. You can set up a transformation from the text masterdata of the original infoobject, to the attribute masterdata of the BPC infoobject to include the id's in the attribute table. If you do so, do not forget to select only id's for one language to avoid duplicate loads.

Former Member
0 Kudos

Has anyone working with scenario 1 run into trouble with the BPC_SYSADMIN id having visibility to the unix directory where the open hub deposits the file?

We're trying several ideas but have not yet found where the issue is, including making the directory permissions wide open.

We get a 'cannot find document/directory' message back when trying to preview the file in the DM package screen.

Thanks for any ideas.

-Amy

Former Member
0 Kudos

When trying to run the dimension lock program ZUJA_UPDATE_BPC_DIM_LOCK I keep getting an error message :'wrong AppSet or Dim". Iu2019ve tried every possible value for AppSet (technical, name, etc). It wonu2019t run. The short dump thatu2019s happening indicates a wrong AppSet name. Anything I am entirely missing here?

0 Kudos

Not for Claudia as I know she is past this however check table UJE_USER.....appset and user should match these entries. Hope this may help others. Cheers

Former Member
0 Kudos

Does you application set name includes lowercase letters ? If so, check my post above...

Former Member
0 Kudos

Thanks

Edited by: Vishal Mahawadhi on Sep 17, 2009 10:43 AM

Former Member
0 Kudos

Hey vishal,

I was able to resolve all those unexpected errors.

Thank you.