Taking one row to write multiple rows in distinct files

[Anonymous]

Hi all - 

I think this is my first net new post, so go easy on me.

My situation is this: I'm working on a proof of concept that takes Salesforce data, transforms it, and then needs to write certain sets of columns on separate lines of an export file. 

So, functionally, if we were to imagine only one row of Salesforce data, the process I need to build is this: 

• tSalesforceInput ->
• tMap ->
• tFileOutputDelimited ->
  • Row 1: Columns A, B, C
  • Row 2: Columns D, E, F
  • Row 3: Columns G, H, I

Basically each row has a separate schema, which, it turns out, is important. 

It's a little more elaborate than that but that's the basic gist.

My issues is that once I've retrieved the salesforce data, cleaned it, and passed it through the mapper to correct formats for the output file, I CANNOT figure out how to write multiple lines. 

I have tried:

1. tSplitRows - didn't work because even when I got to setting up the columns as Field1, Field2, Field3 and the rows defining what should be on the separate lines, each of my lines is a different length. So Talend wants to insert extra delimiters for blank fields, which causes my export file to fail validation. 
2. tUnite - Doesn't work because even when I separate the mapped exports into their separate row schemas and ship the data to temp files or (just discovered) hashes, each of those objects have different schemas and tUnite doesn't like that. (I did, however, learn about the importance of launching things as separate subJobs.)
3. Various iterations of attempting to make tFlowToIterate work. 

I saw one comment here link to an article that I think was written by @rhall (or @rhall ?) that sounds like it solved this using a tJava component, but the website is defunct ( https://www.rilhia.com/quicktips/quick-tip-row-multiplication ). 

My next problem will be generating separate files per row, but I really don't even care about that at this point. I just need the one row -> separate lines deal to work for now.

I'm really at wit's end, here, so any assistance (particularly tonight! The POC is time-critical, unfortunately!) would be massively appreciated. 

Hope you all are staying safe in these weird times. 

[TRF]

Instead of long description, you should include examples of what you have, what you want.
Anyway, maybe tFileOutputMSPositional component could help you in this case.

[Anonymous]

Hello - 

Here's a screenshot of the status quo, pardon the "in medias res" nature of it:

[TRF]

Strange design.
tLogCatcher subjob should be disconnected from others.
On the other side, 2nd subjob should be connected to the 1rst one by a OnSubjobOk trigger (or it may start before the 1rst one finishes).
Finaly, consider tFileOutputMSPositional as suggested in my previous answer.