DSXchange

Hi,

I came across this interesting requirement. Pardon me if this has already been discussed but I couldn't find it anywhere.

I have a source table with columns:

Name, Age, Location

The Location column can contain any value and is not fixed. For e.g. It may contain New York. it may also contain US. So it can be a city, state, country or even a district.

The requirement is simply to load records belonging to different locations to different files. So if I have 20 locations, the job would have to determine the number of values in Location at runtime and create as many files.

I have devised a logic to implement this using a UNIX script and a looped sequencer. But I was wondering if there's an easier way to implement this.

Help Appreciated

Off the top of my head, a Server job (or Server Shared Container) writing to the Folder stage or using the UNIX csplit function after the file has been written could do this.

I'm not sure if I'd want to use server components in my parallel job. Will look into the UNIX command. Thanks you!

In the immortal words of Blue Oyster Cult: Don't Fear the Server.

ha...i am not yet a seasoned campaigner to justify the presence of a server component in my jobs!

Can you explain the logic in simple steps how you implmented the above requirement by using Looping and Unix?

1. Create a single job with a source connector that extracts data for one location. This condition should be parametrized (e.g. WHERE LOCATION=#jp_Location_Nm#).

2. Implement your logic and create a file where again the file's name should be parametrized (e.g. File Name= Target_#jp_Location_Nm#).

3. Now create a looped sequencer with one UNIX script out of the loop. The UNIX script should execute a query against the DB to extract a unique list of all locations and store it in a file (delimited)

4. The file created in the above step gets passed in the Start Loop activity which will pick up one location at a time from the file and pass it to the parameter #jp_Location_Nm#.

5. The sequencer loop gets executed as many times as there are locations in the file and a separate file gets created for each location.

I know that this logic sounds complicated but is actually pretty easy to implement. There might be an easier way out but I can't think of any.