DSXchange

Your job design file definitions don't match the content of the file.

You might want to read this discussion about hash files:

http://www.dsxchange.com/viewtopic.php?t=85364

Look in the online manual for the transformer stage. You can setup a sequential file as a second output from a transformer, where the first output is an ODBC/OCI stage. Checking the reject box in the transformer constraint dialog automatically refers DB rejected rows down that link.

Mona, this topic is extensively covered on the forum. Please use the search facility and enter "oracle stored procedures" and check the radio button "Search for all terms". You will get a full history of issues and clarifications, as you will discover your question is ambigous because "it depends on...

Yes, but I shudder to think of the performance of using a READU lock on every request for the next sequence value. I prefer my own design, and would suggest you consider something similar. Instead of every request obtaining a lock, get a block of numbers and only lock when that block of numbers is e...

With release 5.1 a new daemon was added, the "Hash Cache Daemon". You can run a command that requests the daemon to put a hash file into a shared cache. All jobs that use preload to memory for a hash file will use the hash cache daemon data first before loading its own copy. When you're done, you "d...

Search the forum for mkdbfile command

1. We are looking for the 'best' way of moving our most time consuming jobs to PX. For example, it would be extremely handy to retain the transformer icons etc. But it looks like there is no way to 'move' a job to retain some of the coding. We have looked at exporting the XML and manipulating it so...

Roy, you seem to verify what I was suggesting is going on. Because @FALSE is the constraint, the output link never formats the data and writes it out. This causes your cpu utilization to double (50% --> 100%) and your rows/sec to drop (2300 --> 1300 rows/sec). Sorry, only worked with Tru-64 once 3 y...

For a walk down memory lane....

http://www.dsxchange.com/viewtopic.php? ... ss&start=0

Please post the logic.

I'm not even going to ask why you're putting 67 million rows into a hash file. I'm not even going to get on my soap box and talk about no read/write caching available, inefficiencies, etc. However, if you're going to use a single job to load this hash file, let's say from a sequential file, you're b...

You might consider putting some event logging into your stored procedure to see what it is doing, how far it is getting, etc. I think between Vincent and I we would put money on a bet that your stored procedure is doing something for which you have not accounted. It's probably in la-la land, or busy...

I thought of splitting the read to several jobs but this needs some thougth, since the data is transactions which I need to sum up acording to relevant owner which the transaction belongs to; since there is no partition splitting ownership and the fact that there is no defined number of rows per ow...

Contact your management and convince them to bring in a seasoned DS consultant (checkout the Inner Circle members on this forum) to guide your company in a 1 week consulting engagement. If you're asking, then you need more help than you know. It's not just how to move source code around, it's how yo...