Search found 29 matches

by gabrielac
Wed Sep 17, 2008 10:21 am
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Aggregator stage
Replies: 5
Views: 3593

Order by key columns and date, and use the Remove Duplicates stage. You can choose to keep the first or last row, which would be the same a min or max date.
HTH,
Gaby
by gabrielac
Fri Sep 12, 2008 7:18 am
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Trap Duplicate record
Replies: 7
Views: 3377

If I understood correctly the problem, I would divide it into two jobs.
1. Create Output 2, using remove duplicates.
2. Using Output 2 in a lookup, and create Output 1, with leftover records.
HTH,
Gaby
by gabrielac
Fri Sep 12, 2008 7:08 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Parallel loading with SQL loader Direct path
Replies: 2
Views: 2370

I believe your problem is that Oracle does not allow multiple simultaneous LOADs to the same table at the same time.
HTH,
Gaby
by gabrielac
Fri Sep 12, 2008 7:04 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Compile error
Replies: 4
Views: 2204

Are you getting this error when compiling a specific sequence, or for all the sequence jobs you have? We had a similar problem, when trying to compile any sequence. The error disappeared when the Administrator gave our users Modify permissions on the "Program Files\Ascential" folder HTH, G...
by gabrielac
Thu Sep 04, 2008 6:30 am
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: How to implement "Not In" function in Filter Stage
Replies: 4
Views: 4894

One possibility would be to create a source file with all the possible values, then use a lookup stage, and filter stage to filter those that were not found.

HTH, Gaby
by gabrielac
Tue Sep 02, 2008 6:07 am
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Deadlock detected while inserting into Oracle table
Replies: 6
Views: 5018

- Have you tried using the Load option instead of insert?
- Have you changed the upsert time for commit / rows for commit environment variables?

Gaby
by gabrielac
Tue Aug 19, 2008 7:47 am
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Commit frequency when using Oracle Load/Append
Replies: 15
Views: 11269

The ORA_UPSERT_COMMIT parameters are specific to using the Upsert option, not the Load option. When using the Load option, one temporary file is created for each partition that is running for the Oracle stage. When the job finishes running for all the source rows, then the sqlloader is called for ea...
by gabrielac
Mon Aug 11, 2008 12:54 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: generate osh
Replies: 3
Views: 2805

Have you checked the 'Generated OSH visible ... ' at the Parallel tab in Datastage Administrator?
by gabrielac
Fri Aug 08, 2008 12:46 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Oracle Deadlocks
Replies: 3
Views: 3352

Add the following environment variables as job parameters:

APT_ORAUPSERT_COMMIT_TIME_INTERVAL
APT_ORAUPSERT_COMMIT_ROW_INTERVAL

time = 60
and rows = 2000

worked for me

HTH, Gaby
by gabrielac
Fri Aug 08, 2008 12:25 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Best approach for determining changes
Replies: 6
Views: 4007

It depends on the hardware of the Oracle DB and the Datastage Server, as well as the workload on both at the moment the job is ran. In our case, with similar volumes, it was quicker letting Datastage do all the work.
by gabrielac
Fri Aug 08, 2008 12:03 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Oracle Deadlocks
Replies: 3
Views: 3352

A couple of thoughts that might help:

- Have you tried using the Load option instead of insert?
- Are there other Oracle stages reading from the same table in the same job, such as a sparse lookup?
- Have you changed the upsert time for commit / rows for commit environment variables?

Gaby
by gabrielac
Fri Aug 08, 2008 10:03 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Date Format
Replies: 7
Views: 2983

Try

Trim(Oconv(iconv( .... ), "DYMD[4,2,2]"), " ", "A")

Regards,
Gaby
by gabrielac
Fri Aug 08, 2008 9:10 am
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: SAP BAPI stage query
Replies: 1
Views: 1293

I'm having the same scenario. I have tried BAPI -> File sucessfully with import parameters, but unsuccesfully when I need to complete a table as Input, and another table as Export

Did you encounter the same problem?

Thanks,
Gaby
by gabrielac
Mon Feb 27, 2006 8:29 am
Forum:
Topic: Reporting DataStage Operational Metadata via MetaStage
Replies: 2
Views: 2503

Re: Reporting DataStage Operational Metadata via MetaStage

For starters, your question is a totally different topic than the one discussed here.
On the other hand, Data Lineage / Impact Analysis and the browser are quite different from each other.
by gabrielac
Thu Feb 23, 2006 7:05 am
Forum:
Topic: Sync between Import Category & User-Defined Category
Replies: 4
Views: 1883

You wouldn't need to drag and drop again Entity1 to the user-defined category. However, you still have to send the changes to DataStage.

Gaby