DSXchange

Source: 10gR2 Target: 10gR2 Pulling about 210K rows 54K get rejected to seq file 156K get loaded to target Total load time takes 10 min. Rows/S is around 270. I've tried multiple array size settings, tuned query on source, but I cannot get any better performance than that. I've worked with other ETL...

The 98 avg length is comes from the table stats value I get from the analyze in Oracle. So, I'm sure there are records above that length.

Thanks for the information. I have the same thought about the 64bit and I'm trying that today. I took the default settings on the other hash file and it dies as I've stated above. Post later with the results. [quote="ray.wurlod"]It may be that you have mainly oversized records - these have...

Thanks for the reply. I made sure I had a clean create hash set and re-ran the job. Watched the hash file size grow. The DATA.30 grew to 1.07gig and then stopped. Only the OVER.30 grew from that point and got to 824m and then the job died with: ds_uvput() - Write failed for record id Am I hitting a ...

I'm having a hash file issue. I have a Oracle table that is being loaded with 12,411,177 records. The avg record length is 98. When I use the hash file create within the hash stage to create a dynamic hash file, the job fails after wrting about 10m records to the file. The error I get is: LOAD_IODSU...

[quote="vcannadevula"]Where are you creating the hash files, is it in the UV account or in a custom defined path??[/quote]
I am using a custom defined path to store my hash file.I am using open funtion to read the file.Should i used openpath function instead?

[quote="ray.wurlod"]That message (dimensioned array not found) typically means that the routine has not been compiled/cataloged, or that you have mis-spelled or mis-cased its name where you have used it. Or, if you have used it in a routine, that you have omitted the DEFFUN declaration for...

[quote="vcannadevula"]Where are you creating the hash files, is it in the UV account or in a custom defined path??[/quote]
Its created on a specific custome defined folder

This inbuilt routine is unable to 'see' my hash table when i try and do a test in Manager, it returns "Table Not Found"
When called in Designer, job compiles with an error "Array 'UtilityHashLookup' never dimensioned".Any ideas?

I think this is an old topic, but I'm having issues here I need help with. I read the quote below and my DataStage rejects do not act this way. I'm on v7.5. I have an input data file into a transformer. I also have an Oracle target table and a sequential file for the reject path. I set the reject &q...

This may be an issue with Windows XP SP2. Try going to http://www.asential.com/eservice/pages/ ... tches.html

It's is called DS Client Patch for Windows XP SP2. Apply the patch to each client PC and try again.

[quote="ray.wurlod"]And the other thing is, this being a sharing place, is that we share our solutions. Can you please report how you solved it?[/quote] I was able to do this by using ExecSH Before/After-job subroutine that looked like this: echo "execute pdw_exec_status.exec_status_m...

I found the issue. It was on the select query side. I'm a dummy. But I still think conversions should be easier in DS.

Anyway, thank you very much for your information and time.

[quote="ArndW"]Johm73, If you are doing a string to integer conversion, the function NUM(YourString) will return a 1 if it is numeric, a 0 if it is not. You can use either the FMT() or the OCONV() function to expictly format your output. What is your target Oracle datatype? If your insert ...

Thanks... That was very informative. However, I didn't get an answer out of it. Even if DS does "on-the-fly" conversions when using functions against field values, it must have some datatype to it...or my insert into Oracle would not be failing for trying to insert a character into a numer...