DSXchange

You do not have to specify formatting with Iconv; it's very clever. But you must get the day-month-year order correct. Oconv(Iconv(TheDate, "DDMY"), "D/MDY[Z,2,2]") You may find Date Conversion Demystified an interesting summary. Beware, too, of the CENTURYPIVOT r...

A variable defined in COMMON is only available in one process. In version 5.x and 6.x, two Transformer stages joined by a link would run in a single process (DSD.StageRun), that invoked the subroutine for each in turn. My guess is that you have enabled row buffering in your 7.x version of the job, w...

Is there an easy way to detect changes in the source databases? If so, then DataStage can run periodically to detect these and process those rows very easily. For example, the SELECT statement generated by DataStage could include a constraint on the LAST_UPDATED_DATE column selecting only rows updat...

The SQL is obviously of the form SELECT MAX( column ) FROM TABLE; You can create a DataStage job that collects this value, and writes it directly into the SDKSequences hashed file (the one used by the SDK Key Management routines) where the key value is defined as a constant in the Transformer stage....

Provide you've not enabled transaction groupings, the SQL statements executed on the two input links to an ODBC stage are independent of each other. You can prove this quite easily by enabling ODBC tracing then poring over the results, that very thick book the ODBC API manual in hand. Of course, if ...

You will need to create a custom routine to build the full message body that you've specified, which can then invoke DSSendMail to send the email. The Notification Activity is quite limited - indeed it's just a fancy front end to the DSSendMail function; replace the Notification Activity with a Rout...

Is the listener running?

Validate only is good immediately after deploying into production, as an extra check that DSNs, pathnames and passwords are all good. Unfortunately, since this can't be a parameter, it means deploying other than in read-only mode. I have my own mechanisms (a routine that mimics the way Version Contr...

In the given design, the name of the job executed by Activity1 is hard coded in Activity1 and could therefore just as easily be hard coded (as a string parameter value) in Activity3.

1. Can you document what you want to achieve (not what you want to do with DataStage; what you want to achieve )? 2. Job parameter references need a trailing "#" character; your samples lack this. That is, you specify D:\test\#DataFile when you should be specifying D:\test\#DataFile# to use a job pa...

It is possible, as dhiraj says, but only when the reference input link is fed from the ODBC and UniVerse stage types.
Open Help from Manager or Designer and open the topic called Defining Multirow Lookup for Reference Inputs

Luckily for you, LEN(@NULL) returns zero!

Arguably, given that the job runs OK in the development environment, the problem isn't in DataStage. You're going to need to put on your detective hat and aggressively determine anything that is different between the two systems. Don't take "nothing" as an answer from anyone. I'd be looking at thing...

My solution to the "manual" approach was cited earlier: Of course, you need to ensure that any manually-run job as well as any automatically-run job updates DW_BATCH_CONTROL when it starts and when it ends. I normally ensure this with standard before-job and after-job subroutines. Typically these ro...

There is a Tutorial that comes with DataStage. If you work through this you will obtain at least a fundamental understanding of how to design and execute DataStage jobs.