Using multiple EIM tables in single process section

Hi ,
Often we come across a situation where in some required columns are mapped in one EIM table and some mapped in the other EIM table mapped to a given base table.
Is there a way wherein we can utilize multiple EIM table having mapping of a base table under the same process section with only base table parameter poiting to the target table.?
Regards,
Ashutosh

Hi,
May be I have misinterpreted what has been mentioned in the EIM admin guide.In the page 41 chapter EIM Configuration File the following has been mentioned for the TABLE paramter.Therefore i was wondering if we can use more than one table in a given process section in IFB file.
[TABLE           Required. Specifies the name of an EIM table used in this process.
                     *Multiple TABLE parameters may be used to define a process using*
                     *more than one table*.
                     Example:
                     TYPE = EXPORT
                     BATCH = 101
                     TABLE = EIM_ACCOUNT
                     EXPORT MATCHES = S_ORG_EXT, (NAME > A)
NOTE: For performance reasons, you should limit the number of
tables to export or merge in a single process section to five tables or
fewer.]
Warm Regards,
Ashutosh

Similar Messages

BC4J Problem using multiple EntityObjects in a single ViewObject

Hi,
I have found a bug while using multiple EntityObjects in a single ViewObject.
Considere the following example:
Table A:
A_ID
Table B:
B_ID
A_ID (FK) NULLABLE
TABLE C:
C_ID
B_ID (FK) NULLABLE
For each table there is a corresponding EntityObject : AEntity, BEntity and CEntity.
Now building a ViewObject based on CEntity, BEntity and AEntity, where AEntity and BEntity are referenced via their corresponding associtaions, the following problem occurs:
1. As long as both IDs are not NULL everything works fine, but when I set the B_ID attribute of CEntity to NULL I receive a NullPointerException with the detail "null".
2. If the attribute A_ID of BEntity is NULL the values based on AEntity are not updated at all (they keep the value they had for the last row).
Any information or feedback on this issue would be very welcome!
Regards
Frank

Here's a sample code that works for a
LineItem->OrdView->CustomerView case,
public void setOrdId(Number value)
    //set the boolean if there's a current order.
    boolean hasOrd = (getEntity(1) != null);
    setAttributeInternal(ORDID, value);
    //after the OrdID is set, check if there's a valid Ord Entity in this row.
    if (hasOrd && getEntity(1) == null)
      //if not, then set the Customer Entity to null as well.
      super.setEntity(2, null);
      LinesViewImpl vo = (LinesViewImpl)getViewObject();
      //And force a RowUpdated event for "customer-entity-usage-attributes".
      vo.notifyRowUpdated(findRowSetForRow(null), new oracle.jbo.Row[] {this}, new int[]{6,7});
} Note that in the above, you have to "override" notifyRowUpdated method in the ViewObjectImpl subclass, so that it's available to the LineViewRowImpl subclass. You can do this
globally by creating a custom subclass of ViewObjectImpl that all the VOs in your application "extends".

Multiple DTP's in single process chain

Hello BW guru's
can we have many DTP's in a single process chain? If so, how can we do that?

Krishna,
You can have multipl DTP's in single process chain.
Assume a scenario where you need to update from DSO to Multiple targets.
If you want to run the DTP one after another use AND
otherwise drag and drop the DTP and link it...
Cheer's
HVR.

How to use multiple hierarchies for a single char in single query

Hi,
Is there any way that we can use multiple hierarchies for a single char in single query. I tried and it just allows me to select one hierarchy even if I use hierarchy variable.
I have a requirement where user wants to see information related to a cost center with different cost center groups in different hierarchies (every year has different cost center group hierarchies).
Suppose I want to see information related to a cost center from year 2001-2004.in these four year cost center may have been associated to different groups depending upon that year hierarchy. How can I do that?
Thanks
Jona

Nope. Now way to do this.
There is always just one hierarchy assigned to a characteristic. And even if the hierarchy was time dependent, it only reads it for one key date and not according to transaction data.
Regards,
Beat

How to use multiple tape drive from single client when I want to backup single filesystem?

Hello All.
I want to backup 1 filesystem using 4 tape drives for reducing backup time.
In case of symantec veritas netbackup, they can user "NEW_STREAM" for multi-streaming backup.
Please anybody let me know that how to do use multiple tape drive for single client.
1. Backup Source server : Linux (1 client)
2. Backup Soucre : /data1 (500GB)
-> Current directory structure doesn't have sub directory under /data1, just files present under /data1
If direcotry struceture has a sub-direcotry like( /data1/aaa, /data1/bbb ), Is it availble to use multiple tape drives?
3. Tape Drive with OSB : LTO6 * 4 drives
Thank you.

You would have to create a different dataset for each sub-folder. If you only have files at the top level folder then even the NEW_STREAM option couldn't be used to split the job.
Organise it into folders and then create datasets in a client folder such as :
/usr/local/oracle/backup/admin/config/dataset/Linux/data1_aaa
/usr/local/oracle/backup/admin/config/dataset/Linux/data1_bbb
Then in the schedule you just specify the Linux folder. In each dataset you list the hostname and the folder name, such as :
include host Linux
include path /data1/aaa
Now it will create a new job for each dataset and therefore each folder.
Thanks
Rich

How to use multiple tape drive for single filesystem on single client.

Hello All.
I want to backup single filesystem with multiple tape drive.
Incase of Symantec veritas netbackup "NEW_STREAM" for multi-stream backup.
Backup client has a single backup filesystem like /data1 directory with 4 LTO6 drives.
/data1 directory doesn't have a sub-direcotry, just present a files.
If /data1 directory has a sub-directory like /data1/aaa, /data1/bbb, /data1/ccc , Is it possible to use multiple drives?
Please anybody answer the questions.
Thanks advanced.

See how to use multiple tape drive from single client when I want to backup single filesystem?

Using multiple ipods on a single itunes

I currently have a 40GB ipod and have bought for my wife and daughter 2 2gb ipod nanos'. i have all our music on itunes in one library, how can i download only the songs my wife or daughter want onto their ipods as opposed to downloading every item which is what happens with my ipod currently
Hope you can help
Thanks
Dave

There are a couple of recommended methods for using more than one iPod on a single computer. Method one is to have individual Mac or Windows user accounts which by definition would give you completely separate libraries. Each account has it's own iTunes folder, Library and iTunes Music folder and you load it with CDs etc just as you did with your original one. Each iPod can be set to update however the owner chooses, sync all, manual or sync specific playlists.
Method two is to set your preferences so that either one or all iPods get updated with only certain playlists within one library. Have a look at this article and see what you think and go for whichever you feel suits your needs best: How To Use Multiple iPods with One Computer
Another option when using a single library is to set the iPods to manual update: Managing content manually on iPod
Choosing the update option "automatically update selected playlists only" (called Sync Music - Selected playlists in iTunes 7) allows you to create a playlist specifically for each iPod and drag the tracks you want into it. If you tire of the list and want to change it, you just add or remove the songs you don't want. The ones you take out out remain in the library to be used by the other iPods. Make your playlist a Smart playlist and limit the size to just below the advertised capacity of your iPod ( for example, around 3700MB for a 4GB or 1800MB for a 2GB Mini or Nano). You can read more about playlists at these links:
iTunes: Creating playlists of your favorite songs
How to create a Smart Playlist with iTunes

Use multiple LDB in a single program

Hi,
I want to use multiple log Database in a single program, is there a way to acheive the same..
Thanx.
Arjun.G

Hi,
You can use function module "LDB_PROCESS" to call a second Logical Database from a program where you already assigned one Logical DB.
There is a documentation and sample program available in every SAP system.
1. Call transaction ABAPDOCU
2. Expand node: Database Access > Logical Databases
3. Double click on node: Calling LDBs using a Function Module
There you will see an example how this can be achieved.
<<text removed>>
Regards,
Ogeday
Edited by: Matt on Feb 17, 2009 5:54 AM - please do not ask for points

Using a 2-D array Single Process Shared Variable w/ RT FIFO for comm between a Deterministic and non-deterministic loop on an RT Target

Our problem is that we currently use a 2D array to store CAN data on a Real-time Target. The array is 20 elements of 3 byte elements as so:
                0              1              2
0              [byte]   [byte]   [byte]
19           [byte]   [byte]   [byte]
These values are passed between a Deterministic Timed (DT) loop where they are set and a Non-Deterministic Timed (NDT) loop where they are read and passed into a Network Published Shared Variable (NPSV) for communication across the network to a Host PC. I have insrted an image for illustration, pardon the size.
Currently to pass the data between the DT and NDT loop we are using a Global Variable (GV). To improve the system we have attempted to replace these GVs with Single Process Shared Variables (SPSV) with an RT FIFO enabled.
To create the shared variable I simply right clicked the GV of interest and selected create Shared Variable Node form the drop downs. At this point LabVIEW presented me with a 2D NPSV within a new Library hosted on the RT Target. I then selected this new NPSV from the Project, changed it to a SPSV, and enabled a single element FIFO. This variable was initialized with a default value for the size described above and then used in our code for the DT to NDT communication, and conversion to a corresponding NPSV for sending to the Host.
When I went to run the code I noticed that the variable was in fact 2D, however its size was only 2 elements of three bytes each, in other words only two of the row indices were populated and the other appeared as uninitialized. in addition, this data had no resemblance to the set initilazation value. This was also how the variable was presented on the host side of the network after tranfer into a NPSV.
The peculiar part is that If I change this SPSV to a NPSV and then try to change it back, I receive an error saying the type is not supported for SPSV with an RT FIFO enabled. I have to disable the FIFO (which defeats the entire purpose) in order to successfully compile! I am unclear as to what is the bug in this case. Should I not be allowed to create the original 2D SPSV with a single element RT FIFO enabled without receiving an error? Or if this is okay how do I fix the problems associated with the variable after being allowed to create it?
I have found the following discussion in which a user states “The only limitations for custom controls is the ability to use it with RT FIFO enabled on a network-published shared variable”. Is this also true for SPSV? I have not found any documentation explicitely stating this for SPSV, though it is stated for the NPSVs.

Martin,
RT FIFOs don't support Multi-Dimensional Arrays, which would corroborate the issues you're seeing. You can break up the 2D array into 1D arrays by reshaping the array, then you'll be able to use the RT FIFO enabled variable, just set the array size to the total number of elements (20*3 = 60).
You can also pass the 2D array via pre-allocated queue, or using a Functional Global. We have a reference example for a circular buffer using Functional Globals here.

How to use multiple aggregate functions in single query

hi to all
The output will be giving first_name,last_name,max(salary),min(salary) and the output will be department wise.From employees table in single query
output will be:
first_name|last_name|max(salary)|min(salary)

SELECT first_name||' '||last_name as ename,
       MIN(sal) KEEP (DENSE_RANK FIRST ORDER BY sal) OVER (PARTITION BY deptno) "Lowest",
       MAX(sal) KEEP (DENSE_RANK LAST ORDER BY sal) OVER (PARTITION BY deptno) "Highest"
FROM   emp
ORDER BY deptno, sal;Edited by: Ramio on Jan 10, 2012 10:43 PM

Using multiple physical tables in a single logical dimension table

I have two physical tables that are related on a 1 to 1 basis based on a natural key. One of these tables is already part of my RPD file (actually, it is the W_EMPLOYEE_D from the Oracle BI Applications). The second table contains additional employee attributes from a custom table added to the data warehouse. Unfortunately, I don't seem to be able to display ANY data from this newly added custom table! I'm running on OBIEE 11.1.1.6.
Here's what I've tried to do. Lets call the original table E1 and the new one E2. E1 is part of the repository already, and has functioned perfectly for years.
- In my physical model, I have imported E2 and defined the join between E1 and E2.
- In my logical table for E1, I've mapped E2 to E1 (E2 appears as a source), set up an INNER JOIN in the joins section for E1 and added the attributes from E2 in the folder
- In the SOURCES for this logical table, I've set the logical level of the content for E2 appropriately (detail level, same as E1)
- In my presentation folder for E1, I've copied in the attributes from E2 that were included in my logical table
Consistency check runs smoothly, no warnings or errors. Note: E2 contains hundreds of rows, all of which have matching records in E1.
Now, when I create an analysis that includes only an attribute sourced from E2, I get a single row returned, with a NULL value. If I create an analysis that includes one attrribute from E1 and one from E2, I get all the valid E1 records with data showing, but with NULL for the E2 attributes. Remember, I have an inner join, which means that the query is "seeing" E2 data, it is just choosing not to show it to me! Additionally, I can grab the query from the NQQuery.log file - when I run this SQL in SQL*Developer, I get PERFECT results - both E1 and E2 attributes show up in the SQL - the query engine is generating valid SQL. The log file does not indicate there are any errors either; it does show the correct number of rows being added to cache. If I create a report that includes attributes from E1, E2 and associated fact metrics I get similar results. The reports seem to run fine, but all my E2 attributes are NULL in Answers. I've verified basics, like data types, etc. and when I "Query Related Objects" in the repository, everything looks consistent across all 3 layers and all objects. E2 is located in the same (Oracle) database and schema as E1, and there are no security constraints in effect.
I've experimented with a lot of different things without success, but I expected that the above configuration should have worked. Note that I cannot set up E2 as a new separate dimension, as it does not contain the key value used to join to the facts, nor do the facts contain the natural key that is in both E1 and E2.
Sorry for the long post - just trying to head off some of the questions you might have.
Any ideas welcomed! Many thanks!
Eric

Hi Eric,
I would like you to re-check on the content level settings here as they are the primary causes of this kind of behavior. You could notice that the same information might have written down in the logical plan of the query too.
Also, as per your description
"In the SOURCES for this logical table, I've set the logical level of the content for E2 appropriately (detail level, same as E1)"
I would like to check on this point again, as if you had mapped E2 to E1 in the same logical source with an inner join, you would get to set the content level at E1 levels themselves but not E2 (Now, that E2 would become a part of the E1 hierarchy too). This might be the reason, the BI Server is choosing to elimiate(null) the values from E2 too (even you could see them in the sql client)
Hope this helps.
Thank you,
Dhar

HOW CAN I USE MULTIPLE INNERJOINS IN A SINGLE SELECT STATEMENT?

HI,
I AM SHABEER AHMED,
I AM GETTING AN ERROR WHILE I ATTEMPT TO EXECUTE A SELECT STATEMENT WITH MULTIPLE INNER JOINS . BECOZ I WANT TO FETCH ITEM DATA, PARTNER DATA BASED ON HEADER DATA .
THEN OF COURSE I HAVE FETCH DATA FROM VBAK VBAP VBKD SO LZ SEND ME THE SOLUTION.
BYE

Hi,
1.Just see this:
SELECT * INTO CORRESPONDING FIELD OF TABLE itab
FROM t1 INNER JOIN t2 ON t1f4 EQ t2f4
INNER JOIN t3 ON t2f5 EQ t3f5 AND
t2f6 EQ t3f6 AND
t2f7 EQ t3f7.
2.But better to use for all entries.It increases the performance.
FOR ALL ENTRIES
Tabular Conditions
The WHERE clause of the SELECT statement has a special variant that allows you to derive conditions from the lines and columns of an internal table:
SELECT ... FOR ALL ENTRIES IN <itab> WHERE <cond> ...
<cond> may be formulated as described above. If you specify a field of the internal table <itab> as an operand in a condition, you address all lines of the internal table. The comparison is then performed for each line of the internal table. For each line, the system selects the lines from the database table that satisfy the condition. The result set of the SELECT statement is the union of the individual selections for each line of the internal table. Duplicate lines are automatically eliminated from the result set. If <itab> is empty, the addition FOR ALL ENTRIES is disregarded, and all entries are read.
The internal table <itab> must have a structured line type, and each field that occurs in the condition <cond> must be compatible with the column of the database with which it is compared. Do not use the operators LIKE, BETWEEN, and IN in comparisons using internal table fields. You may not use the ORDER BY clause in the same SELECT statement.
You can use the option FOR ALL ENTRIES to replace nested select loops by operations on internal tables. This can significantly improve the performance for large sets of selected data.
Example for ALL ENTRIES
DATA: TAB_SPFLI TYPE TABLE OF SPFLI,
TAB_SFLIGHT TYPE SORTED TABLE OF SFLIGHT
WITH UNIQUE KEY TABLE LINE,
WA LIKE LINE OF TAB_SFLIGHT.
SELECT CARRID CONNID
INTO CORRESPONDING FIELDS OF TABLE TAB_SPFLI
FROM SPFLI
WHERE CITYFROM = 'NEW YORK'.
SELECT CARRID CONNID FLDATE
INTO CORRESPONDING FIELDS OF TABLE TAB_SFLIGHT
FROM SFLIGHT
FOR ALL ENTRIES IN TAB_SPFLI
WHERE CARRID = TAB_SPFLI-CARRID AND
CONNID = TAB_SPFLI-CONNID.
LOOP AT TAB_SFLIGHT INTO WA.
AT NEW CONNID.
WRITE: / WA-CARRID, WA-CONNID.
ENDAT.
WRITE: / WA-FLDATE.
ENDLOOP.
INNER JOINS
In a relational database, you normally need to read data simultaneously from more than one database table into an application program. You can read from more than one table in a single SELECT statement, such that the data in the tables all has to meet the same conditions, using the following join expression:
SELECT...
FROM <tab> INNER JOIN <dbtab> AS <alias> ON <cond> <options>
where <dbtab> is a single database table and <tab> is either a table or another join expression. The database tables can be specified statically or dynamically as described above. You may also use aliases. You can enclose each join expression in parentheses. The INNER addition is optional.
A join expression links each line of <tab> with the lines in <dbtab> that meet the condition <cond>. This means that there is always one or more lines from the right-hand table that is linked to each line from the left-hand table by the join. If <dbtab> does not contain any lines that meet the condition <cond>, the line from <tab> is not included in the selection.
The syntax of the <cond> condition is like that of the WHERE clause, although individual comparisons can only be linked using AND. Furthermore, each comparison must contain a column from the right-hand table <dbtab>. It does not matter on which side of the comparison it occurs. For the column names in the comparison, you can use the same names that occur in the SELECT clause, to differentiate columns from different database tables that have the same names.
The comparisons in the condition <cond> can appear in the WHERE clause instead of the ON clause, since both clauses are applied equally to the temporary table containing all of the lines resulting from the join. However, each join must contain at least one comparison in the condition <cond>.
Example for INNER JOINS
REPORT demo_select_inner_join.
DATA: BEGIN OF wa,
carrid TYPE spfli-carrid,
connid TYPE spfli-connid,
fldate TYPE sflight-fldate,
bookid TYPE sbook-bookid,
END OF wa,
itab LIKE SORTED TABLE OF wa
WITH UNIQUE KEY carrid connid fldate bookid.
SELECT pcarrid pconnid ffldate bbookid
INTO CORRESPONDING FIELDS OF TABLE itab
FROM ( ( spfli AS p
INNER JOIN sflight AS f ON pcarrid = fcarrid AND
pconnid = fconnid )
INNER JOIN sbook AS b ON bcarrid = fcarrid AND
bconnid = fconnid AND
bfldate = ffldate )
WHERE p~cityfrom = 'FRANKFURT' AND
p~cityto = 'NEW YORK' AND
fseatsmax > fseatsocc.
LOOP AT itab INTO wa.
AT NEW fldate.
WRITE: / wa-carrid, wa-connid, wa-fldate.
ENDAT.
WRITE / wa-bookid.
ENDLOOP.
Regards,
Shiva Kumar(Reward if helpful).

Creating XML transformation using multiple internal tables

Hi everyone, 
 
I'm trying to transforme 3 internal tables (from customer master data) into a single XML document. 

DATA: BEGIN OF wtab OCCURS 0 , 
 kunnr LIKE kna1-kunnr, "Customer ID 
 ktokd LIKE kna1-ktokd, 
 land1 LIKE kna1-land1, 
 name1 LIKE kna1-name1, 
 ort01 LIKE kna1-ort01, 
 pstlz LIKE kna1-pstlz, 
 spras_iso LIKE kna1-spras, 
 smtp_addr LIKE adr6-smtp_addr, 
 stras LIKE kna1-stras, 
 END OF wtab. 
 
DATA: BEGIN OF wtab_o OCCURS 0 , 
 kunnr LIKE knvv-kunnr, "Customer ID 
 vkorg LIKE knvv-vkorg, "Sales organisation 
 waers LIKE knvv-waers, 
 END OF wtab_o. 
 
DATA: BEGIN OF wtab_p OCCURS 0 , 
 kunnr LIKE knvp-kunnr, "Customer ID 
 vkorg LIKE knvp-vkorg, "Sales organisation 
 parvw LIKE knvp-parvw, 
 kunn2 LIKE knvp-kunnr, 
 END OF wtab_p. 

The internal tables are related to each other as follows:* 
 
wtab-kunnr = wtab_o-kunnr 
 
AND 
 
wtab_o-vkorg = wtab_p-vkorg 
wtab_o-kunnr = wtab_p-kunnr 
 
I couldn't figure out how to declare this relationship when calling the transformation. Is it possible? 

ABAP 
 
 
REFRESH : gt_source_itab. 
CLEAR : g_rxml. 
 
GET REFERENCE OF wtab INTO gs_source_wa-value. 
gs_source_wa-name = 'DEBMAS04'. 
APPEND gs_source_wa TO gt_source_itab. 
 
GET REFERENCE OF wtab_o INTO gs_source_wa-value. 
gs_source_wa-name = 'E1KNVVM'. 
APPEND gs_source_wa TO gt_source_itab. 
 
GET REFERENCE OF wtab_p INTO gs_source_wa-value. 
gs_source_wa-name = 'E1KNVPM'. 
APPEND gs_source_wa TO gt_source_itab. 
 
TRY. 
 CALL TRANSFORMATION Z_XSLT_CLIENT 
 SOURCE (gt_source_itab) 
 RESULT XML g_rxml 
 OPTIONS xml_header = 'without_encoding'. 
CATCH cx_root INTO gs_rif_ex. 
 
 gs_var_text = gs_rif_ex->get_text( ). 
 MESSAGE gs_var_text TYPE 'E'. 
 ENDTRY. 
 
 
Transformation Z_XSLT_CLIENT: 
 
 
<xsl:transform version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform" xmlns:sap="http://www.sap.com/sapxsl"> 
<xsl:output indent="yes" encoding="UTF-16" method="xml" version="1.0"/> 
<xsl:strip-space elements="*"/> 
<xsl:template match="/"> 
<DEBMAS04> 
<xsl:apply-templates select="//DEBMAS04/item"/> 
</DEBMAS04> 
</xsl:template> 
<xsl:template match="DEBMAS04/item"> 
 <IDOC> 
 <xsl:attribute name="BEGIN">1</xsl:attribute> 
 <E1KNA1M> 
 <xsl:attribute name="SEGMENT">1</xsl:attribute> 
 <KUNNR> 
 <xsl:value-of select="KUNNR"/> 
 </KUNNR> 
 <KTOKD> 
 <xsl:value-of select="KTOKD"/> 
 </KTOKD> 
 <LAND1> 
 <xsl:value-of select="LAND1"/> 
 </LAND1> 
 <NAME1> 
 <xsl:value-of select="NAME1"/> 
 </NAME1> 
 <ORT01> 
 <xsl:value-of select="ORT01"/> 
 </ORT01> 
 <PSTLZ> 
 <xsl:value-of select="PSTLZ"/> 
 </PSTLZ> 
 <SPRAS_ISO> 
 <xsl:value-of select="SPRAS_ISO"/> 
 </SPRAS_ISO> 
 <SMTP_ADDR> 
 <xsl:value-of select="SMTP_ADDR"/> 
 </SMTP_ADDR> 
 <STRAS> 
 <xsl:value-of select="STRAS"/> 
 </STRAS> 
 <xsl:apply-templates select="//E1KNVVM/item"/> 
 </E1KNA1M> 
</IDOC> 
</xsl:template> 
<xsl:template match="E1KNVVM/item"> 
 <E1KNVVM> 
 <xsl:attribute name="SEGMENT">1</xsl:attribute> 
 <VKORG> 
 <xsl:value-of select="VKORG"/> 
 </VKORG> 
 <WAERS> 
 <xsl:value-of select="WAERS"/> 
 </WAERS> 
 <xsl:apply-templates select="//E1KNVPM/item"/> 
 </E1KNVVM> 
</xsl:template> 
<xsl:template match="E1KNVPM/item"> 
 <E1KNVPM> 
 <xsl:attribute name="SEGMENT">1</xsl:attribute> 
 <PARVW> 
 <xsl:value-of select="PARVW"/> 
 </PARVW> 
 <KUNN2> 
 <xsl:value-of select="KUNN2"/> 
 </KUNN2> 
 </E1KNVPM> 
</xsl:template> 
</xsl:transform> 
 
The way the call transformation is now, for each line in table wtab (DEBMAS04) I'm getting all the lines in table wtab_o (E1KNVVM) and table wtab_p (E1KNVPM) 

 
Thank you for your help, 
 
Giselle 


Hi Giselle,
I think it's possible (using xsl:if for instance), but an easier way is to nest your internal tables declarations (wtab_o inside wtab, kunnr becomes useless, and wtab_p inside wtab_o), and use the SAP standard "ID" transformation (no need for a custom XSL transformation). You'll get a slightly different result, but I don't think you want something very sophisticated.
Sandra

Load into single target table frm multiple source table in single interfac

Hi
I have four source table and a single target table.
I need to move data from either of these tables into a target table , and we have to decide the source table based on user input.
Example :
Lets say there are four tables A,B,C,D and one target table T.
If user input says A
then the data from table A will move to table T
And again , if the user says table C then data from table C will move to table T.
And we have to create only one interface for achieving this in oracle Data Integrator ( ODI ).
You can take assumptions in source and target table.

Hi ,
In ODI 11g , there are new feature callled dataset. It allows to use UNION , MINUS etc.
Google it , you will get many tutorials on Dataset. check the link
http://www.rittmanmead.com/2011/06/odi-11g-new-mapping-and-interface-features-part-1/
In your case , you can provide filter conditions on your tables i.e.
Say My target table is EMPLOYEE , My source tables are EMPLOYEE and DEPARTMENT
INSERT INTO EMPLOYEE(CUSTOMER_ID , CUSTOMER_NAME) SELECT CUSTOMER_ID , CUSTOMER_NAME from employee where 'EMPLOYEE' = :EMP
UNION DEPARTMENT_ID , DEPARTMENT_NAME from departments where 'DEPARTMENT' = :EMP ;
Just pasted the Screenshots on following page : http://oracoholic.blogspot.in/ . Have a look
Edited by: user8427112 on Jan 8, 2013 11:04 AM

Best way to combine multiple fact tables in single mart

Hi, quick question that I think I know the answer to, just wanted to bounce it off everyone here to make sure I'm on the right track.
I have a HR datamart that contains several different fact tables. Some of the facts are additive across time (i.e. compensation - people get paid on different days, when I look at a month I want to see the total of all pay dates within that month). The other type of fact is more "status over a set of time" - i.e. a record saying that I'm employed in job X with a salary of Y from a given start date to a given end date.
For the "status over time" type facts, if I choose January 2009 (month level) in the time dimension, what I'd really like to see is the fact records that were in place "as of" the last day of the month - i.e. all records where the start date is on or before 1/1/2009, and whose end date is on or after 1/1/2009. Note that my time dimension does go down to the day level (so you could look at a person "as of" the middle of the month, etc. if you're browsing on a day-by-day basis)
I've set up the join between the time dimension and the fact table as a complex join in the physical layer, with a clause like "DIM_DATE.DATE >= FACT.START_DATE AND DIM_DATE.DATE <= FACT.END_DATE". This seems to work perfectly at the day level - I have no problems at all finding the proper records for a person as of any given day.
However, I'm not quite sure how to proceed at the month level. My initial thought is:
a) create a new LTS for the fact table at the month level
b) in the new LTS, add the join to the time dimension
c) in the new LTS, add a where clause similar to LAST_DAY_IND = 'Y' (true for the last day of each month).
Is this the proper way to do this?
Thanks in advance!
Scott

Hi Scott,
I think you're on the right track but I don't think you need the last part. Let me generalize the situation to the following tables
DAILY_FACT (
DAILY_FACT_KEY NUMBER, -- PRIMARY KEY
START_DATE_KEY NUMBER, -- FOREIGN KEY TO DATE DIMENSION FOR START DATE
END_DATE_KEY NUMBER, -- FOREIGN KEY TO DATE DIMENSION FOR END DATE
DAILY_VALUE NUMBER); -- FACT MEASURE
MONTHLY_FACT(
MONTHLY_FACT_KEY NUMBER, -- PRIMARY KEY
MONTH_DATE_KEY NUMBER, -- FOREIGN KEY TO DATE DIMENSION, POPULATED WITH THE KEY TO THE LAST DAY OF THE MONTH
MONTHLY_VALUE NUMBER); -- FACT MEASURE at MONTH LEVEL. DATE_KEY is at END of MONTH
DIM_DATE(
DATE_KEY NUMBER,
DATE_VALUE DATE,
DATE_MONTH VARCHAR2(20),
DATE_YEAR NUMBER(4));
DIM_DATE_END (ALIAS OF DIM_DATE for END_DATE_KEY join)
Step 1)
Make the following three joins in the physical layer:
a. DAILY_FACT.START_DATE_KEY = DIM_DATE.DATE_KEY
b. DAILY_FACT.END_DATE_KEY = DIM_DATE_END.DATE_KEY
C. MONTHLY_FACT.DATE_KEY = DIM_DATE.DATE_KEY
Note: The MONTHLY_FACT DATE_KEY is joined to the same instance of the date dimension as the START_DATE_KEY of the DAILY_FACT table. This is because these are the dates you want to make sure are in the same month.
Step 2)
Create a business model and drag DIM_DATE, DAILY_FACT and DIM_DATE_END into it.
Step 3)
Drag the physical table MONTHLY_FACT into the logical table source of the logical table DAILY_FACT.
Step 4)
Set DAILY_VALUE and MONTHLY_VALUE to be aggregates with a "SUM" aggregation function
Step 5)
Drag all required reporting columns to the Presentation layer.
Step 6)
Create your report using the two different measures from the different fact tables.
Step 7)
Filter the report by the Month that joined to the Start Date/Monthly Date (not the one that joined to the end date).
Step 8)
You're done.
The act of combining the two facts into one logical table allows you to report on them at the same time. The strategy of joining the START_DATE_KEY and the MONTH_DATE_KEY allows you to make sure that the daily measure start date will be in the same month as the monthly fact table.
Hope that helps!
-Joe
Edited by: Joe Bertram on Jan 5, 2010 6:29 PM

Using multiple EIM tables in single process section

Similar Messages

Maybe you are looking for