DSXchange

I am not at a PX computer or have access to the documentation, but the route I would take is: DateFromDaysSince(-1,StringToDate(DateToString(CurrentDate(),'%yyyy%mm%dd')[1,6]:'01')) You might have to weed out the syntax errors and/or function names yourself, but the b...

If you run this with a single node configuration file what happens?

I don't know, when I was doing support for this product I used "very busy" to be synonymous for "I can't be bothered" But that was long before IBM took over so I'm sure things have changed for the better and they now have echelons of expert support personell slaving away on our issues.

I've now asked 3 times and I am about to give up. How are you determining that the data is missing? View-Data in the designer? LIST command from the ">" prompt? Dump to a sequential file? Are you certain that these records were written to the hashed file (i.e. replace the Hashed file stage with a se...

TAX_CODE_FULL looks unique. You haven't answered the question as to how you know the rows are missing. Have you tried a COUNT or the output-to-sequential file?

If you actually multiply the 2 numbers together you will get the correct answer (try writing a routine with the code Ans = "8.0000" * "21.4300" and you can see that DS does the math correctly. So your rounding is happening somewhere outside of that statement. What is your target ...

What have you declared as your key in the hashed file? And how did you discover that the records are missing? I see no reason, if TXN_CODE is your only key column, that any records are not written to the hashed file. Can you write a simple job that reads the hashed file and writes it straight to a s...

According to IBM support they are "very busy" and haven't even tried to reproduce the problem yet! Is there anyone out there running HP-UX and PX that might run a test job (which I would send via e-mail) for me?

The key field(s) of a hashed file may not contain the @FM (char 254) character.
Could you please explain exactly what your problem is - are records being dropped, do you get errors, is column information missing?

Remove the single quotes from around the variable, i.e.


p.s. Welcome to DSXChange, but please use more meaningful subject lines in the future.

ArndW wrote:You can also define multiple writers per node for the output.

My mistake, I was thinking of multiple readers and then multiplue outputs to a fileset.

You can also define multiple writers per node for the output.

My first guess now would be tthat one or more ORA_ environment settings have changed, particularly since all jobs using load are affected.

How and in which stage are you doing the rounding or conversion by moving the decimal over 3 places? The answer to this will go a long way to answering your question

This is getting odder and odder... Do you have other jobs that use the LOAD method to Oracle that are working without this warning? Might you have changed any of the environment variables related to Oracle?