DSXchange

If it helps any more, I just put into play a job that has a hashed file lookup involving over 20 million rows, a total of 1.2GB, and it works fine. We will need to monitor this and perhaps convert to 64-bit addressing if the data volume grows too much, but for now it's all good.

Or you can call such a stored procedure from an ODBC Enterprise stage.

You create the state file (assuming that's the key generate mechanism you propose to use) with a job that has the Surrogate Key Generator stage with no links. You initialize the state file in a job that has the Surrogate Key Generator stage with an input link, that presumably receives the result of ...

We'll also need to know the data type of each column and stage variable mentioned in the code.

What do you mean by "invalid"? A parameter value that is invalid based on data type - for example "FRED" as a date - will not even allow the job to start. Otherwise, use Arnd's offered solution, but limit the link to 1 record because 0 means "unlimited".

You can test whether the catalog entry exists with a query such as executed from the Administrator client.

You have not really provided enough information. In short, however, use the appropriate stage type for inserting rows into the table, and figure out how you will capture the error code and description.

Example: A job that uses a Surrogate Key Generator stage to initialize a state file has no links and therefore has only one stage.

Because sequential processing is what you get on one node; a stage (operator) can create only one process on one processing node (unless it's a composite operator, which possibility was not canvassed in the question). Pipeline parallelism is between stages (operators), not within one stage, therefor...

You don't connect to a project any more.

You connect to IBM Information Server (machinename:9080) and authenticate there. That having been done, you will be presented with a drop-down list of DataStage server/project combinations, and can choose one from that list.

It can be gotten to work on Windows XP Professional only, but you're on your own; no-one will support it, since Windows XP is not an officially supported operating system.

Looks like a question from a (practice?) certification exam to me.

1. Yes.
2. Yes, depending on the row size in the hashed file.
3. There are no limits if your hardware will support what you seek to do.

If you have a UNIX emulator (such as MKS Toolkit) you can use the find utility, executing the DataStage UVfile command, to identify hashed files. If you prefer a more "pure" Windows approach, you can use the DataStage uvwalk utility to traverse your tree structure (and perform a UVfile command). You...

You're in trouble if "they" ever change the values, Amar. That's why the mnemonics such as DSJS.RUNOK are to be preferred.