What is the “Integration” check box for in the workflow Start task?
Assuming you are referring to the Start task of an asynchronous workflow, when this property is selected, the workflow is used to migrate data from staging tables in IBM TRIRIGA records. This type of workflow is used extensively in IBM TRIRIGA DataConnect.
Check out this IBM Knowledge Center topic about DataConnect that describes the “Integration” check box: Workflow task settings.
[Admin: The same question is also posted in the main Application Platform forum. To see other related posts, use the Staging tag or DataConnect tag.]
In my current project, there was a suggestion to extract (updated) data from TRIRIGA, with a high frequency, and import it into some kind of data warehouse (DW) or business intelligence (BI) solution. Then, from there, perform more advanced reporting and analytics. Have other TRIRIGA solutions implemented something similar? Are there any TRIRIGA best practices or recommendations for staging area, extract-transform-load (ETL), DW, or BI reporting solutions?
[Admin: This post is related to the 12.15.16 post about the IBM TRIRIGA Connector for Watson Analytics. To see other related posts, use the ETL tag or Analytics tag.]
I have created an object migration (OM) with its workflow. The execution works well on the Development environment. But after an import of the OM package with all objects needed, the execution didn’t work on the Test environment. The object migration launch works. The triIntegration workflow launch works. The execution of the request works in SQL Server. The connection in my object migration works.
But there is no row in the staging table S_CSTPHINTERMARCHECONTRAT. Also, I see in the logs:
Calling SQL: [INSERT INTO S_CSTPHINTERMARCHECONTRAT(DC_JOB_NUMBER, DC_CID, DC_SEQUENCE_ID, DC_STATE, DC_ACTION, DC_GUI_NAME, TRIIDTX, CSSTPHHPIDRATTTX, CSTPHRETIRETX) VALUES (?,?,?,?,?,?,?,?,?)] with params[402, 0, 1, 1, 4, cstPHInterMarcheContrat, 2013/M0166, 101GT, ]
I found the problem. The configuration of the integration object was for the Development environment and not for the Test environment.
[Admin: To see other related posts, use the DataConnect tag.]
I have 2 questions regarding the use of DataConnect (DC) and was hoping someone may have the answers I’m looking for:
- 1. Is there a recommended maximum file size for importing data into TRIRIGA using DataConnect via staging tables?
- 2. We have a two-stage DataConnect process where we first create a record (using one integration object) and then associate locations to the created record (using a separate integration object). We split this into separate IO records due to the parallel processing of DC trying to create things in the wrong order. However, when our update record runs, it overwrites the values that are blank in the staging table, even though we are not directly mapping them in the workflow. Is there any way to avoid this?
We are encountering some issues running DataConnect on a DB2 database and wondered if anyone could point us in the right direction. Previously, we were successfully running on an Oracle Database, but since moving to DB2, we have come across some issues.
We found some notes that said we could resolve some issues by ensuring that numeric fields in the import file contained zeros instead of blanks, but we still get the same issues. The import of the data to the staging table is failing, and the TRIRIGA integration object is displaying errors. For some reason, it is not recognizing the data in the file…
You may find an issue with the setup of an integration object record when using a business object that exists with the same name in two different modules. For example, when setting up the Data Map, you can select the Module as triItem and the Business Object as triChecklistCategory. But then you don’t get any options in the Form drop-down (just a blank value), so you are unable to continue in setting up the Data Map. You may observe that there is a triChecklistCategory BO in both the triItem and Classification modules.
The issue is that integration object logic to display business objects with staging tables in the Data Map drop-down list did not take into consideration that the same-named business object can belong to multiple modules. This was reported in TRIRIGA 3.5.1/10.5.1.
As suggested, the issue was that the SQL retrieving the BO staging tables was not looking at the module. Thus, if two BOs with the same name belonging to different modules had staging tables, the integration object logic did not know which BO to use. The fix is to pass in the module name into the SQL. Moving forward, we resolved an integration object “File to DC” issue, where staging tables were not correctly loading in the Data Map tab, if the business object selected had the same name as another business object belonging to a different module.
I have a DataConnect (DC) task (triOption) that worked before, but has stopped working and throws a stack trace. The main DC task runs. But this is a sub DC task. The BO is published. I was advised to check if the necessary UOMs (units of measure) are in the staging table.
So my dumb question is: Does it mean every NU (number) that contains a UOM needs a corresponding UOM column? For example, TRISUBTOTALNU, TRISUBTOTALNU_UOM, and TRIRATENU would need TRIRATENU_UOM?
There seems to be a problem with your metadata, and probably not your DC staging table metadata, but maybe with the triOption’s fields metadata. I tried to reproduce this issue and was unable to reproduce it with various numbers and UOM fields. As for your questions…