DSXchange

dvpawankumar - there are so many contexts that just using our own wouldn't make sense. Generally, hashed files are preferred over sequential files when you need to access the records randomly for reading. Hashed files are analogous to database tables for this example, so just think of reasons why yo...

mystuff - could you post the SQL that is causing the error message you listed?

you can trigger an export from your client PC using "dscmdexport", and you can add that call to your PC's scheduling tool (probably AT)

What do you mean by "getting automatically unscheduled"? Does this happen immediately, or after a couple of executions or days? Since DataStage doesn't have its own scheduler you are probably using cron on your system - have you looked at the cron tabs or checked to see if someone is modifying thing...

The amount of storage required depends upon several factors, starting with the amount of data you have and then depending upon what you are doing with this data - if you pass the data through and just do transformations the amount of temp storage can be as small as 0. If you store all your data in i...

When the hashed file was not populated, how many rows went from the OCI stage to the hashed file? If "0" then you need to look at your database and SELECT statement.

This routine and job use inordinate amounts of CPU time to perform a relatively simple action. I would recommend re-writing it from scratch using somewhat different logic. Pseudocode recommendation: Loop execute shell to return "exists(control file)" true or false IF not(exists...

How jobs are run using Enterprise Edition depends upon how the configuration files are set up. Unless you configure specifically for distributed processing jobs written on a SMP box but run on another system and using the same configuration files will run just the same way.

What happens when you run this from your favorite query tool? Does it return a (correct) result or the same error?

Yoga, it seems that you are actually talking about 2 distinct migrations here, a) Migrating from DataStage V3.x server to DataStage 7.5.x server b) Converting from Server jobs to Parallel jobs. Step (a) might not be as difficult as you assume, whereas step (b) is going to be time-consuming. There is...

The named pipe mechanism requires that you have both a writer process and a reader process running at the same time. You have timeout or wait attributes that control how much time you have available before you get a "broken pipe". So a question would be "are you certain you are running the 2 jobs at...

In your sequence, put in a call to "SLEEP RND(5)+5" to give you a minimum 5 second delay with a random additional delay. This might be enough to keep your server from overloading while starting up all those processes.

ds_ram, Ray already suggested you check to see if you have security software and if so, to see if it is blocking the DataStage port. Have you checked with your system administrator? Also, have you used the default port or chosen another, in which case you need to specify the port in the datastage cl...

Datasets are not magically faster than sequential files. In this case with a 2 node configuration the overhead for splitting the data might be just enough to end up being the same speed as a single threaded sequential file. You should try a controlled test. Write a job with a row generator to write ...

Hello Bhaggu and welcome to DSXchange. This question has been posed and answered many time in the past. If you use the SEARCH facility you will find a lot of threads on this subject. If you have a specific situation which isn't covered by the other answers, you would need to explain in a bit more de...