Search found 53125 matches

by ray.wurlod
Fri Dec 19, 2003 5:39 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Yet another issue with dates!!!
Replies: 10
Views: 1636

Can you use CURRENT_DATE, or SYSDATE, or whatever it is, in your SQL or load script?
If that's the wrong data type, you're deep in the brown stuff!
Or, similar, make the default value for that column CURRENT_DATE or whatever.
by ray.wurlod
Fri Dec 19, 2003 5:34 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: days in month
Replies: 8
Views: 1707

The "for Dummies" series is published by a company that owns the rights to the title concept. However, I would argue against "DataStage for Dummies". It is really too complex a product for this kind of treatment. All you could cover is how to drive the tools. Allow me to share one of my favourite qu...
by ray.wurlod
Fri Dec 19, 2003 5:29 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Manuals and Software
Replies: 6
Views: 1134

Jane works out of Singapore, but does cover India. Peter Zeglis recently came on board for Ascential to manage education and professional services in Asia-Pacific; he is based in Sydney but, like all of us, is on the road a lot. He was in Kuala Lumpur yesterday, for example. I posted earlier that he...
by ray.wurlod
Fri Dec 19, 2003 5:16 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: fuction for coverting a character string to numeric value ?
Replies: 3
Views: 936

I suspect the question is about determining the ASCII value of a character. The function for this is SEQ() - provided that you only want to process single-byte characters. For example, SEQ("A") returns 65, and CHAR(65) returns "A". If you need to process multi-byte characters, then the functions you...
by ray.wurlod
Fri Dec 19, 2003 5:10 pm
Forum: Data Integration
Topic: The ETL Staging Database (Sandbox) concept is not dead yet
Replies: 7
Views: 5488

Getting buy-in for the staging area isn't too difficult give clients who are prepared to listen intelligently to rational argument. In essence there must be two (minimum) weapons in the consultant's armoury here; a few horror stories about where it went pear-shaped, and why (none of these examples h...
by ray.wurlod
Thu Dec 18, 2003 11:46 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: days in month
Replies: 8
Views: 1707

Sandyha originally sent me a private message. When I replied to that I highlighted the advantage of making questions available to more potential answerers, and of sharing those answers, by posting on the Forum. As a new poster Sandhya may not have gotten as far as the FAQ. Maybe be we need a "read m...
by ray.wurlod
Thu Dec 18, 2003 4:54 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: compare 2 hash files
Replies: 4
Views: 1297

... or even a query (for example, for one side of the difference):

Code: Select all

SELECT hashedfile1 WITH EVAL "@RECORD" <> EVAL "TRANS('hashedfile2', @ID, -1, 'X')"
by ray.wurlod
Thu Dec 18, 2003 4:42 am
Forum:
Topic: Informix stage
Replies: 1
Views: 1248

For extraction, VERY similar, provided the database and client software and connectivity components are properly set up (like I said, VERY similar). There are a few things that are syntacticaly different for user-defined SQL; Oracle, for example, has an idiosyncratic way to represent an outer join i...
by ray.wurlod
Thu Dec 18, 2003 12:12 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: compare 2 hash files
Replies: 4
Views: 1297

Is there any way to do this with an Oracle table? A hashed file is a table in the UniVerse database (aka the DataStage Engine database). If the answer is no in one, it's no in the other. If you want to access rows in a hashed file, you use the correct query languages, just like you would with other ...
by ray.wurlod
Wed Dec 17, 2003 6:06 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Checking for files on NT from UNIX
Replies: 19
Views: 2976

chulett wrote:the specs came to me on Monday and it goes into "production" tomorrow. Ha!

Bets? :o
by ray.wurlod
Wed Dec 17, 2003 4:13 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: NT to UNIX
Replies: 10
Views: 2209

Very easy. Export from one, import on the other.
About the only things you need to look out for are:
    ODBC driver setup
    pathnames (less of a problem if parameterized)
    calls to ExecDOS (replace with ExecSH, fix commands)
    calls to DSExecute() with Shell argument of "NT"
by ray.wurlod
Wed Dec 17, 2003 3:49 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: NOTFOUND link variable curious behaviour
Replies: 10
Views: 5440

IsNull() is appropriate on the key value for the lookup (since keys must be NOT NULL) but not on non-key columns, which might return null even if their row is found. Personally I have never had a problem with the NOTFOUND link variable, but all (well, at least 99%) of my lookups are to local hashed ...
by ray.wurlod
Wed Dec 17, 2003 3:42 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: copying rows is it an issue
Replies: 4
Views: 865

It's a design issue. You have to make sure that your logic allows the last row to be processed (yours currently doesn't, unless the trigger column changes value in the final input row). Or you need to append a dummy row to the input file. Search this forum for ideas on how to do that.
by ray.wurlod
Wed Dec 17, 2003 3:40 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Sql*Loader option PRESERVE BLANKS
Replies: 5
Views: 3200

Which version of the ORABULK stage are you using? Actually it won't matter very much, all versions are very old, and not written to the DataStage plug-in API. In fact the first version was written in DataStage BASIC - may still be (I haven't checked lately). It probably is due for an overhaul; Ascen...
by ray.wurlod
Wed Dec 17, 2003 3:33 pm
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: APT_ParallelSortMergeOperator: Error
Replies: 12
Views: 5488

Code: Select all

SELECT NAME, JOBNO FROM DS_JOBS WHERE NAME = 'name_of_job';


The JOBNO gives you a job number. The directory you need is RT_BPnn where nn is the job number.