DSXchange

If you create a transform that calls a routine there is no difference. A transform is simply code that is inserted in-line to the generated program. If you can code your logic in a single line (without calling a routine) it will execute more efficiently as a transform. But, since it is limited to a ...

I agree that if all that your Transform does is call your routines, then there is going to be no speed difference. It was only on Wednesday I was listening to a UK Ascential Software person saying to avoid Routines in favour of Transforms (that is with the code in the Transform). Of course this is n...

Is that true your mail system and DS server should on same box? how do they communicate each other. I tried before it did not work for my sutiation: DS server, OS win2000, box1; mail server, MS exchange system, box2; DS client, OSwin2000, my desktop. Thanks! -----Original Message----- From: Raymond ...

I dont really want to start any kind of holy war of Routines versus Transforms, but the note below reminded me of something Ive been wondering about, off and on, for a long time. When I first started working with DataStage (back when the documentation still said VMark, if that means anything to anyo...

Hello, I have a datastage job which does a full table scan on a staging area table with 125 columns. There are about 15 hash lookups in that job. The maximum amount of rows in a hash being 600000 for one. I am getting a performance of 130 rows per second. I have set my array size to 10000 and may tr...

David, I totally agree with you about using hash file look up. I was trying send out my question below to everyone yesterday, somehow, it didnt sent out. My question is related to Hash, maybe you can help me a little bit. ****************************** I was trying to take advantage of using merge s...

You need to bear in mind that any routine you use will slow down the job as it will be called each time a row is passed through a job. If you can get away with using a Transform(s), however this may mean having several different ones to suit your different dates, this is quicker when running the job...

PHB??? It this some Aussie term that I presume is not very complementary? -----Original Message----- From: Raymond Wurlod [mailto:wurlodr@au1.ibm.com] Sent: Thursday, March 28, 2002 4:30 PM To: datastage-users@oliver.com Subject: Re: how to determine if SQLServer ODBC native drivers are used I would...

I would strongly suggest that you ask the DBA. Clearly, on UNIX, you are not using Microsoft-supplied ODBC drivers. This is for the simple reason that Microsoft do not supply drivers for UNIX platforms. If your DBA cannot accept this, then suggest the job acronym be changed to PHB. Other than that, ...

Were running DataStage 4.1 on AIX, a client of WIN2000, with the normal target being Oracle 8.1.6. If we design a job with an ODBC source stage, and the ODBC connection is to a SQLServer2000 database across the network, the SQLServer DBA wants to be sure we use the SQLServer ODBC native drivers. How...

I had a similar situation: passing down parameter based on an input record. For example, the input record value is a system date,(or whatever date you assigned) and you want to look up a table of last week run records with #START-DATE# AND #END-DATE#. First,you figure out firstday, lastday in DS env...

You can write a datastage routine in Datastage Manager and use it in the transformation. This way all your validation is done in routine and can be called from many jobs if needed. Palanisamy Sakthivel Participant Datastage Newgroup -----Original Message----- From: Castelein, Maaike [mailto:maaike.c...

Hi, Not living in a perfect world I have problems with sequential input files containing bad data : for example dates in different formats (I want to reformat all dates into one format) or just plain wrong (2002-02-31). Does anyone have practical information on how to clean and validate your data in...

make sure you make a copy of your dsenv file in the DataStage server engine directory, before upgrading. I have seen it overwritten before. ----- Original Message ----- From: "Rui Soares" To: "Datastage-Users (E-mail)" Sent: Wednesday, March 27, 2002 10:45 PM Subject: Upgrade DS to 5.1 - Solaris > >...

With reference to number 2 - DataStage buffers the writes to a hashed file, so even if the monitor showed all records as being processed DataStage might still be writing to disk. You should only see this with a large number of records. ----- Original Message ----- From: "Sameh Bishara" To: Sent: Wed...