DSXchange

Running more jobs on the same machine without increasing cpus means things will run slower. What's the problem? Are you running any diagnostic tools to show you cpu, memory, and disk utilization?

Create a table with the same column definitions as your load data and enable for parallel query. Bulk load your data into the work table using DIRECT path option and TRUNCATE method. Create an index on your update key (should be unique, hopefully you're not doing wildcard updates). Analyze your tabl...

There are multiple ways to use multiple cpus in the same job. 1. Have independent streams of logic. Ex: seq1a--->xfm1--->seq1b seq2a--->xfm2--->seq2b This will cause xfm1 and xfm2 to operate independently as streams and use 2 cpus. 2. Have chained active stages and enable inter-process row buffering...

In a real time scenario, an SCD can become a critical timing issue. Imagine you have a queue of database transactions waiting to be processed. This queue contains rows of data that feed both dimension and fact tables. If the rows are time sequenced according to their updated time, then you can feel ...

To execute anything within your directory you need to make sure your PATH has "./" in it. Setup your user profile.

Make sure that the permissions to this script include read/execute access to your ds_scripts directory.

ravikiran2712 wrote:Thanks for your reply Ray,
Can i Use MS SQL SERVER load for this purpose. Any help would be appreciated.
Thanks,
Ravi

BCP is the "MS SQL SERVER load" utility.

I think the r for descending order goes after a key's notation:

Do you mean with a Unix sort command or the Sort stage? For the command line, specify the sort key using byte notation and sort order along with data type for left/right sorting. For the Sort stage (really just a wrapper over the Unix sort command), create two columns that represent the data for eac...

Haven't looked lately, but what's wrong with command line BCP bulk loading a sequential text file? If you don't like the BCP stage, you can do it the old fashioned way and write your own script to load the data file.

There's no documented API to check if a job is compiled or not. The best you can do is check its run state.

There are also undocumented APIs that can be used to programmatically compile a job. You can search this forum, they are posted in multiple places.

Column definitions are irrelevant unless you're using the UV/ODBC stage on the hash file. Hash files are content agnostic, there are no table definitions unless you want the hash file to be treated as a table. If so, your standard SQL DDL will create a hash file with a relational layer over it to al...

Can you tnsping or do anything as the specific DataStage user? It seems the error message is a permissions issue to access the Oracle installation directories for the client software.

When you installed the Oracle client, did you do it as Administrator and make it available to all users?

No, you can't do that.

Actually, as long as the jobs are not trying to clear the hash file while others are writing to it, you're fine to have multiple jobs streaming output to the same hash file. You'll have to be aware that the same row being written to the file by different jobs will not be the best design, because you...