Hello All,
I have been developing in Ab Initio for about 7 years and I am now evaluating Talend. One problem I am having is that I want to use Talend to generate a set of test files (account, customer, acct_x_cust, transactions) that have referential integrity. I will be using these files for subsequent testing.
The problem d'hour is that I have generated my account records and am trying to generate my acct_x_cust (cross reference) data, in which I use the generated acct_id's and generate new cust_id's. Each acct_x_cust record has a role code, and each account can have multiple customers tied to it. To create the acct_x_cust, I have used a tMap with three outputs (one for each customer role code). The trouble now is merging these three outputs into a single acct_x_cust file. tUnite does not want to do it. I remember reading that two outputs from tMap cannot go to the same component). I have tried to throw sorts on each flow between the tMap and the tUnite, but it's not taking it.
What are my other options here? How can I merge three flows with the same metadata into a common output?
I am willing to concede that using a different tool, perhaps I need to shift my mindset, so if anyone has a suggestion as to another method of accomplishing the same goal, I am all ears. The important thing is creating a dataset *with* referential integrity.
Thanks!
Shane
Hello,
You're right, currently, Talend do not allow any cycle.
In most of the case, you will probably not need to have 3 different output from the same tMap with the same schema.
Can you post a screen shot of your tMap for a precise answer ?
Regards,
That is what I was afraid of... I have attached a screen cap of my tMap. I want to create 'secondary' customers from some accounts and 'triciary' customers for some accounts, but they all share the same account number. Hopefully that makes sense. Perhaps I need to approach it from another angle.
Thanks,
Shane
Hi
Because Talend do not allow any cycle, so you can't link the three file to tUnite component directly.
To solve this issue, you can take the ways:
1) Copy and paste the input components with the same properties, so you can link the different outputs to tUnite component.(see screenshot 1)
2) Output the three outputs to three temporary files, then, read the data from these temporary and link to a tUnite component.(see screenshot 2)
Best regards
shong
I don't think the normalize in that example would work, because it would give me the same customer number for roles 2 & 3. Perhaps I could generate the number of roles, then based on that generate X amount of customer id's and normalize those... I will give that a shot. I guess I would have to code that in the tMap in some kind of 'for loop'? Is that supported? Where would I find documentation on the syntax?
Thanks,
Shane
Hi Because Talend do not allow any cycle, so you can't link the three file to tUnite component directly. To solve this issue, you can take the ways: 1) Copy and paste the input components with the same properties, so you can link the different outputs to tUnite component.(see screenshot 1) 2) Output the three outputs to three temporary files, then, read the data from these temporary and link to a tUnite component.(see screenshot 2) Best regards shong
Shong,
Could you explain what you mean by 'cycle'. It's a new term for me. I understand that Talend does not allow you to unite flows that originate from the same component, but I really don't understand why. I don't understand why the components are not completely independent. To me a flow, is a flow, is a flow. It may have characteristics, obviously metadata, whether it is a continuous flow (iterate?), but it does not care where it comes from, or where it is going... in my mixed up mind at least.
Thanks!
Shane
