DSXchange

I usually write to a standard name and change it in an after-stage or after-job subroutine. To use @USER0 you would need to craft your own routine, but the logic is straightforward. If, for example, the file is in the current (project) directory, you can use &UFD& as a Type 19 UniVerse file....

Welcome aboard! :D Please check this post about why there is no such thing as "urgent" here. I would be very surprised if SQLTABLES does not exist; it is a system table containing information about all the tables in the database. It's more likely that you don't have SELECT privilege to that and othe...

Two choices.

1. Lookup on the key and test the returned non-key value within stage variables.

2. Create your hashed file with more than one key column, and perform multiple lookups.

Option 1 is the more efficient.

The answer about renaming a project remains "no".

Ascential's web site is probably best.

Do you have transaction logging enabled (is the parameter TXMODE in uvconfig set to 1)? The SQL DELETE statement would probably create too big a transaction in any case; the maximum size of a transaction within DataStage is set by other uvconfig paramters, such as LOGBLSZ and LOGBLNUM. This is moot ...

It's a bug in 7.5x2 - it doesn't work at all for most activities, particularly Job Activity.

DON'T The ORABULK stage has been around since version 1.0, and is quite inefficient. Prefer the OCI stage with Load mode, or a Sequential File stage to write the DAT file. The ORABULK stage does not have capability for automatically starting sqlldr, which means you have to create a sqlldr command in...

Do all three environments point to the same target table (not the same name, the same table)? I suspect not. They shouldn't. Anyway you need to make sure that the tables are identically defined (particularly column nullability) AND that the definition matches the definition used in your DataStage jo...

The code from the Job Control tab that is generated by compiling the Job Sequence is what is meant. Don't forget to wrap in in Code tags.

If you read the Parallel Job Developer's Guide chapter on the Join stage the point that the input links must be sorted on the join keys is explicitly made. Hash partitioning on the same keys guarantees that all join keys will be found on the same processing node where the particular join is being pe...

Also try setting PWD=.

htrisakti3 wrote:Have you tried:

iDate = IConv(fieldName, "D-YMD")

iDate is now integer & you can easily Max(iDate),

But you have to: OConv(iDate, "D-YMD")
to display this back as proper date

Iconv and Oconv are not natively available in parallel jobs.

That may be the case, but someone has typed that in as a parameter reference in a DB2 stage in your job design. Hunt it down and correct it.

Could be. Access violation means an attempt to get a memory address that you don't own. On your PC it could be a memory address that doesn't exist! Add another 2GB of RAM!