DSXchange

I can't seem to find the answer either in the documentation or by using the search facility here, but I think I've done it before... Strings with embedded quotes are usually "doubled" when written to sequential files. This applies to server jobs and such .csv as generated from Excel: Actual string c...

Yes, you can use the stage concurrently. Basically any object that is only being read can be re-used, difficulties only occur when simultaneous reads and writes are attempted.

Hello Rafidwh,

the first task that will help you do this is to use the search facility and use the keyword(s) "upgrade" and "7" and then to read those threads which contain problems & solutions that people have had in doing such an upgrade.

Thanks for the answers - even though they weren't the ones I wanted to hear.

I think that there is a big market for porting ICONV/OCONV to PX. Might even be worth 'productizing'!

ray.wurlod wrote:... and a DSDetachJob to indicate that you're finished with it and another process doesn't get a "locked" message.

You caught me out; I noticed the omission when I posted yesterday and was too lazy to go back and edit the post; even though I knew that you would notice and correct my error.

One of the processes in this job has failed. What does stage QAS_ADDR.QASClean do?

This type of conversion is very common, so common that you can specify it in the sequential read stage of PX.

The Parallel Job Developer Guide that you can read as a .PDF in the Ascential folder of your Client PC has the information you need.

Yes, you can have multiple readers of the same sequential file. Protection is usually N-readers OR 1-Writer.

deepesh, The best way for you to overcome this problem would be to actually read PDF documentation on before and after stage subroutines. The shell script (ksh on Windows?) command invoked did not complete successfully and returned a '1' - the before/after expects a successful call to return '0'. Tr...

You understood incorrectly, DataStage at 7.5 still uses the UniVerse tables to store metadata.

The first issue is not your absolute speed, but that it changes over the run of the program; this points to some issue - once you get that cleared up you can approach your array sizes and commit frequencies.

What was the result of writing to /dev/null instead of the table?

Nagesh, change your job to make the output go to a sequential file with the path "/dev/null" and see what your performance numbers are and if they remain constant. If the speed slows down over time as it did before then you can eliminate that part of your job. If the speed remains constant then you ...

You will have to write some DataStage Basic code to replicate these two functions manually. I took a look at some definitions on the web for this and they can easily be programmed - except the actual formulas that I have seen are somewhat different, so you would have to choose the "best" ones for yo...

The routine would contain:

- a call to DSAttachJob that you wish to reset
- a call to DSGetJobInfo to get the current status
- a call to DSPrepareJob to make the job runnable

Doesn't Oracle expect a '/' instead of a space in the default date format?