DSXchange

It's always worth cleaning entries that are no longer needed from file systems. It's the fact that people don't that keeps disk vendors in business (in part at least). Pretty anything in the directories identified as scratch disk resource is a candidate (provided no jobs are running), as well as any...

What you say is true. It also leads to the danger that the lock table may become full, as noted earlier. This could cause the job to abort. Depending on how the lock table tunables are set, there may be as few as approximately 8000 locks available.

You CAN do whatever will compile successfully. Whether it's best practice is another question entirely. I have a strong belief that production jobs ought never to abort; that the control hierarchy (job sequences) should always remain in control.

Use one Sequential file stage with a filter command that uses cat (UNIX) or type or copy (Windows) to load all the rows from all the files into a single stream. Run this through the Aggregator stage. Even better would be to run the output of cat through sort (on the key) as part of the filter; this ...

Actually the lock is only taken if the lookup fails. Provided you're writing back to the hashed file in every such case, you won't have a problem with the lock table, as the write releases the update lock.

In a word, MORE. You cannot stint for parallel jobs. If you can't get another file system, then you have no possible solution for "file system full" and you should go back to server jobs until you can get some more disk space. Parallel jobs are far more resource hungry than server jobs; the extra re...

It might be the best, but be very, very careful when using it that you don't inadvertently - or even deliberately - release validly-held locks. The locks are there for a reason - to prevent data anomalies in the Repository database.

For anyone comfortable with multi-valued fields (for example in UniVerse or UniData), the Make Vector stage should not be daunting. A vector is simply a list of values in the one field.

If you're on 7.5.x you can click on the Generate Report tool (second from the right) in the Designer toolbar. What exactly were you after?

What Ken said.

Umm.. ask "them" to support Teradata in the Stored Procedure stage?

Sorry about the paucity of other ideas; I tried to be comprehensive the first time, and didn't check whether Teradata was supported for the SP stage.

You can use a UNC pathname with OpenPath provided that the ALLOWNFS configuration tunable is set to 1. You may also be able to use the hostname!pathname convention, but I have no way of checking this at the moment - it may fail, because it will try to trigger UV/Net software (which isn't installed w...

Set your array size to 1 and your rows per transaction to 1. The link variables Arnd mentioned can then return the correct values, because your job is now sending one row at a time. So any returned error information is related specifically to that row.

When you use the expression editor to build the expression, and choose "job parameter" from the menu, you probably note that, in expressions, the surrounding "#" characters are not required. That you used them is probably what made the compiler complain.

Real and dfloat are approximate representations of data. Decimal and numeric are exact representations. They are different data types, which is why you experienced what you did. My original response was based on your statement that the data type was Decimal.