DSXchange

How about write a java, perl, or shell script instead and read the queue and act accordingly. I don't think your design is sound. You're doing job control, using data transformation stages. You'd be better off using a constantly running script, periodically reading the queue and taking the appropria...

Either: 1. Import selected and put a big pot of coffee on to brew while you wait for it to dig thru 400MB of data. When it finally comes up, select all of the routines. 2. Make a copy of the dsx file and open it using Wordpad (DON'T USE NOTEPAD!!!) and highlight everything between BEGIN DSJOB and EN...

Change the name of your output file to something else, then use job control to decide if the file created is worthy of being renamed into the file you desire. In other words, use an after-stage/job routine to rename the file if the file created is more than 0 bytes.

IMO, it's even harder to teach the "art" of hash files than the "science". Without a doubt, every single DS engagement I've been on over the last 7+ years has been impacted by ETL developers ability to understand the underyling concepts of data movement. The added complexity of hash file tuning has ...

First things first. Your ability to supply a result is your responsibility. If you want fast and accurate answers, consider the price you are paying. $0 US. That's pretty cheap. We don't bill for time or answers, we just ask that you don't request URGENCY, because we have lives too. Plus, as is the ...

Have you ever tried indirection? In the Unix world, that's a common method for avoiding exposing command line options to the ps command. Your scheduler might indirect in the encrypted parameter from a hidden
file.

lstsaur wrote:you have to hardcode the user ID and the password in your ISQL command script

Really? A batch, perl, or java script can't fetch these values from a protected parameter file or environment variable when executing the command line call or even better use trusted connections?

Okay, so it's the easiest table/index design possible. Unless you can check the explain plan, we're pretty much shooting ducks at night. Is the size of your row very large, meaning that you have enormous rollback usage? What is your commit rate? If I were your DBA, I would do the whole work table, d...

Forgot to answer directly. Yes, I was trying to say that your partitioned loading sessinos will have to cooperate with the sequence, and that there is an amount of serializing that could take place. Tom Kyte seems to not worry so much about it, but I couldn't find where the topic of parallel partiti...

Consider this thread: m GO about 2/3 of the way down to get the discussions about bulk loading 100's thousands of rows simultaneously. Transactional inserts with sequences isn't a big deal, I agree with Tom 100%. (Buy Kytes Expert-One-On-One, it's an excellent book). I guess my point is that you hav...

Welcome aboard. Please make use of the excellent Search facility and the over 69000 posts, some of which cover this in detail.

Write a script to connect to your database of choice and execute the "do something" stored procedure. You'll have full capability to audit, check, log messages, whatever. The best part of it is that you can run these on Windoze using the Command stage. You also can use the Batch job to call the DS A...

How about putting some DSLogInfo calls into the "Else" conditions to see why some of those (just a guess) are failing. If the else condition is taken and you don't act accordinly and the logic continues, bad things can happen.

NOT THE ENGINE DIRECTORY. Messing with that has disastrous consequences. You need to make sure the project directory itself has the appropriate permissions.

If you have urgent questions please contact Ascential technical support. The answer for storing hash files is that by default they go into your project directory, unless you specify a project name in the account field on the stage properties. The better choice is to use externally pathed hash files ...