DSXchange

Don't think so. This one looks like a C++ compiler issue.

Did you install the C++ compiler (from Visual Studio .NET) before installing DataStage?

It's set to 50 if you run from Designer.

Go to Director, open Options from the Tools menu to set a different limit.

20 million is not too bad. All you need in the hashed file to test for existence is the business key values (and maybe the surrogate key values as a non-key column) from the current rows of the dimension table. Two integers is not such a big row; storing 20 million of these in a hashed file won't ev...

To remember the previous row you need to record the value AFTER performing the comparison.

The row count for a link is given by @OUTROWNUM (for output links from a Transformer). You don't need to calculate anything while the job is running. As noted, you can interrogate the job subsequently. You could also generate 1 as a single column on an output link and pass this through an Aggregator...

Tags, including Code tags, are affected by the check boxes at the bottom when you reply, especially "Disable BBCode in this post".

Click the Format tool in your routine editing window.
Check that the final End statement lines up with the RETURN(Ans).
If not, there's an imbalance.
Post the formatted code.

You have reinvented an intrinsic statement called Locate().

My guess is that the sequence and the table are not synchronized. There's nothing in Oracle to force them to be synchronized. Someone has added rows to the table without using the sequence. You need to re-synchronize (re-initialize) the sequence using SELECT MAX(key)+1 FROM tablename as the value.

Is the job REALLY still running, or has it aborted but not been able to reset its status file (from which Director obtains the status)?

Probably.

Why?

You can purchase the add-on Java PACK.

Or you can purchase the add-on Java PACK.

Or when writing to the hashed file read cache preparatory to performing lookups.

But my comment applies equally to the situation when the hashed file write cache fills.

No, it's just a warning. When you run out of memory for hashed file cache, the hashed file is accessed directly on disk. You don't lose data, but you do lose speed.

Apply a TRIMB() before anything else. Stick the result in a stage variable. Use the Field() function to get the state and zip. For example zip might contain five or nine characters. Beware that the city name may have spaces in it, so you need to know how many fields there are in the string, and extr...