Search found 53125 matches

by ray.wurlod
Thu Nov 10, 2005 2:29 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: geting error while importing schema from orchestrate
Replies: 1
Views: 1274

The message means that the FTP service is either not configured or not enabled on the server. There must be an entry for it in /etc/services, and its port number must not be blocked by a firewall.
by ray.wurlod
Thu Nov 10, 2005 2:27 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: How to implement control break in PX?
Replies: 8
Views: 2797

Have you investigated the Surrogate Key Generator stage? Otherwise you'd design it like you would a server job, substituting a Data Set stage for the Hashed File stage. And you would have to restrict execution to sequential mode, and guarantee that the Data Set is updated before the next row is proc...
by ray.wurlod
Thu Nov 10, 2005 2:19 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: The job has to abort after 5 rows enter into the table.
Replies: 8
Views: 2573

DSLogFatal() does have a side effect of aborting the job, but do you want to check each time a row is written to the table? This will slow things too much. You can add a row limit to the link writing to the rejects table; in the constraints grid you will find an "abort after rows" column. Simply set...
by ray.wurlod
Thu Nov 10, 2005 2:15 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: How do you make the job in runnable state after aborts
Replies: 7
Views: 1310

What you want to do is not possible with a job in isolation.

The "reset if required" functionality can only be performed in some kind of wrapper, such as a job sequence or a shell script.
by ray.wurlod
Thu Nov 10, 2005 2:12 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: MSSql Server Load Stage.
Replies: 2
Views: 825

I guess "they" will tell you that it's because SQL Server only runs on Windows.

Depending on your version of SQL Server you may be able to use the BCP stage.

But "they" have a SQL Server Enterprise stage that can be used in parallel jobs. :roll:
by ray.wurlod
Thu Nov 10, 2005 2:08 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Tracing from a log directory back to the job
Replies: 2
Views: 731

Code: Select all

SELECT NAME, CAST(JOBNO AS INTEGER)
FROM   DS_JOBS
WHERE  NAME NOT LIKE '//%'
ORDER BY 2;
by ray.wurlod
Thu Nov 10, 2005 2:04 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Problems with MSSQL text data type
Replies: 1
Views: 1024

Change the data type of the text field to VarChar 4000 (Long VarChar may still work). Change the derivation of the text field to CAST(textfield AS VARCHAR(4000)).
by ray.wurlod
Thu Nov 10, 2005 2:00 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Can we send email from DS jobs???
Replies: 11
Views: 3223

... or using an operating system command via ExecDOS :wink:
by ray.wurlod
Thu Nov 10, 2005 1:55 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Removing unwanted Parameters From Job
Replies: 4
Views: 1011

Yes, of course, but do you understand the concept of a collection within a "UniVerse" table? If not, it would be a highly dangerous undertaking that would be likely to void your warranty. There are three collections (associations of multi-valued fields with a collection header indicating the number ...
by ray.wurlod
Thu Nov 10, 2005 4:18 am
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: julian date to yyyymmdd conversion in PX jobs
Replies: 10
Views: 7995

I vaguely recall it was JD Edwards that began that weird convention of a single digit century designator. Saves one byte. Whoopee!! They're a pain to deal with - and consume unnecessary CPU cycles in doing so, imho. If you're happy with a BASIC Transformer stage, you can use Iconv() and Oconv(). Oco...
by ray.wurlod
Thu Nov 10, 2005 4:10 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Can we send email from DS jobs???
Replies: 11
Views: 3223

You can use ExecDOS (or ExecSH on UNIX) to invoke an operating system command to send email. You can pass information via job parameters that become command line arguments to your command (BAT file?). You will, of course, need a mail client on the machine where DataStage server is running. That's wh...
by ray.wurlod
Thu Nov 10, 2005 4:07 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: time variant tables
Replies: 6
Views: 1472

If it's really a Type 2 SCD you insert a new row. The only column you update in the old row is expiry date, or whatever other method you might use to indicate that it's an expired row. Indeed, a Type 2 SCD doesn't need any such column; the row with the maximum surrogate key value for a particular bu...
by ray.wurlod
Thu Nov 10, 2005 4:03 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Error Trapping, sort of
Replies: 2
Views: 671

There's no great penalty to having more columns on links. Create a separate flag column for each of the columns you have to check, so that you can pass them downstream for subsequent checking. Saves mucking about with your actual data. Do you need lookups for all 427 columns? Is there no other mecha...
by ray.wurlod
Thu Nov 10, 2005 3:54 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: How to pass job parameters from a file
Replies: 5
Views: 1869

My favourite, now that 7.5 is out, is the Trans() function in an expression, or the $ReturnValue activity variable associated with an upstream activity in a job sequence.
by ray.wurlod
Thu Nov 10, 2005 3:48 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: datastage in a real-time environment?
Replies: 6
Views: 3536

Many large telecommunications companies run server or parallel jobs in "near real time" (small frequent batches) taking rows from mediation machines (maybe thousands per second, and maybe using multi-instance jobs to service multiple mediation machines). Mediation machines are basically between the ...