If you are unable to create a new account, please email support@bspsoftware.com

 

News:

MetaManager - Administrative Tools for IBM Cognos
Pricing starting at $2,100
Download Now    Learn More

Main Menu

Loading CSV File

Started by tamise, 30 Dec 2013 03:12:03 AM

Previous topic - Next topic

tamise

Hello,

I want to use Datamanger to load a csv file.
The problem is that i want to run just only one time per month the job.

So my questions are :
- If i run a second time is there a way to not count this time ?
- Is there a way to run a build only if the csv file exists on the repository (for exemple  for a month there may be not exist a csv file).


Thankss

MFGF

Hi,

I'm not quite sure what you mean by your first question. Do you mean you want the second run to deliver the same results as the first? What refresh type are you using in the build that delivers the data? If you use (for example) Truncate, you will replace the target table with the data the build is delivering this time. Is this what you mean?

If your build is being triggered by a jobstream, you could precede the fact build node in the jobstream with a condition node that uses a filecheck() function to test for the existence of the text file you want to read in. A "false" result could either wait for a designated period then loop around and try again, or else finish the jobstream without running the build. What is your desired outcome?

Cheers!

MF.
Meep!

tamise

Thank you for your answer.

For the first question : I do not want that the job run if it has already run one time for the month.
For the second question: the function  filecheck() has to be used on a procedure node ?

Tks.

MFGF

Quote from: tamise on 31 Dec 2013 06:40:41 AM
Thank you for your answer.

For the first question : I do not want that the job run if it has already run one time for the month.
For the second question: the function  filecheck() has to be used on a procedure node ?

Tks.

How can you check if the job has already run? Are you setting a date/time flag somewhere? If not you could set one each time the job runs, and use a condition node to evaluate whether the build has already run this month before firing it off.

Filecheck() can be used in the expression of a condition node - it seems like the most appropriate thing to do here :)

MF.
Meep!