DSXchange

Why do you want to re-invent the DataStage Director? The API is used to create utilities to be used on the server. In fact, dsjob is provided as an example of a program that uses the API. This is fully documented, with source code, in the manual. Similarly, to be able to run jobs from a client based...

Simplistically, an active stage is one that processes rows and has the capacity to change the data being processed, for example the Transformer, Aggregator, Pivot or Sort stage. A passive stage is one that connects to a persistent data store, either retrieving rows from or putting rows to that store...

You could try user-defined SQL using CALL - I do not know whether this will work. As a general rule, definitions of stored procedures to be invoked by DataStage have to be imported (or created) first (so that the number of arguments can be verified) and must have at least a return result set. In my ...

In DataStage 6.0 you can create PX jobs on a Windows-based server. You can not compile them there, which means that you cannot test them there. You can not use View Data, because the underlying (Orchestrate) engine has not been ported to the Windows environment. I believe that PX on Windows is suppo...

There is a useful technique called a "paragraph" that you can use. You set it up with the Administrator client's Command window, through which you enter each of the commands in turn. Note that the commands are stored in the history pane. You can multi-select commands from the history pane and click ...

A couple of things to add. Check for any use of ExecDOS, which you will need to change to ExecSH. Similarly check for any use of DSExecute, where you may need to change the Shell argument from "NT" to "UNIX". Tip: export the project and search the DSX file for ExecDOS and DSExecute. Remember, too, t...

$DSP.Close is a function called from the stage. SQLFreeHandle encapsulates the ODBC (BCI) functions SQLFreeStmt, SQLFreeConnect and SQLFreeEnv; in this case it appears that it is experiencing difficulty releasing the handle to the connection (hDBC). This is something that only Ascential Engineering ...

Remember that the benefits of using a hashed file technique obtain only if an "=" comparison are done on the entire primary key. You can enhance performance on non "=" comparisons (irrespective of whether you use your original model or one of the subsequently suggested ones) by judicious use of inde...

I have used multiple input links to one OCI stage, and it appeared only to use one database connection.
There was not much else happening at the time, so I can't be 100% sure; it worked so I didn't pursue it! [8D]

Ray Wurlod
Education and Consulting Services
ABN 57 092 448 518

I think you'll find that the DataDirect drivers work only for a one month trial period (from first use), then you have to pay for them. As delivered, they are licensed only for use by DataStage stages, not for use by BCI code. Check out their web site: m Note also that Ascential will NOT support "my...

Is the information in the recent thread on Client Access software any value?

Ray Wurlod
Education and Consulting Services
ABN 57 092 448 518

I have one client still on 4.0.1. Hope that's some comfort.

They plan to move to 6.x "sometime real soon". And from Windows to UNIX. And from one target database to a different one. Whew! [:0]

Ray Wurlod
Education and Consulting Services
ABN 57 092 448 518

Have you examined the job designs to ensure that every column returned from every hashed file is used in the job? By eliminating those that aren't, you get smaller - faster loading - hashed files and better throughput. Similarly, are all the rows in the hashed file necessary? For example if you are ...

Depending on which stage type you're using, you can also take advantage of transaction grouping, which means that DataStage can send update, then insert, then commit.


Ray Wurlod
Education and Consulting Services
ABN 57 092 448 518

To answer the original question, all you have to do is lie. Don't declare Hash.B as a key column (in this job's Transformer stage). However, this introduces the possibility that there is more than one row in the hashed file that meets your criterion (Data.A = Hash.A). A hashed file stage will only r...